One read of Claude Developer Platformapi-20261028T203725Z
116 pages moved out of 821 read.
Pages moved
116
significant first
Pages read
821
in this capture
Captured
20:37 UTC
Corpus hash
82e28663d5c2
corpus-hash
What this read moved
26-50 of 116, page 2 of 5This capture is too large to show at once. Changes 26-50 of 116 are below, significant first; the rest are on the following screens.
build-with-claude/batch-processing Changed · +60 / -10 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling ## Data retention
from line 2
22title: Batch processing
33url: https://platform.claude.com/docs/en/build-with-claude/batch-processing
44description: Process large volumes of Messages requests asynchronously with the Message Batches API, cutting costs by 50 percent and increasing throughput.
5featureMetadata:
6 status: ga
7 zdr:
8 eligibility: not-eligible
9 note: 29-day retention; async storage required.
10 supportedModels:
11 - claude-fable-5-1
12 - claude-mythos-5-1
13 - claude-fable-5
14 - claude-mythos-5
15 - claude-mythos-preview
16 - claude-opus-5-5
17 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
21 - claude-haiku-4-5-20251001
22 - claude-opus-4-8
23 - claude-opus-4-7
24 - claude-opus-4-6
25 - claude-sonnet-4-6
26 - claude-opus-4-5-20251101
27 - claude-sonnet-4-5-20250929
28 supportedPlatforms:
29 Claude API: ga
30 Claude Platform on AWS: ga
31 Claude Models on Google Cloud (Anthropic managed): ga
532---
633
734Batch processing is a powerful approach for handling large volumes of requests efficiently. Instead of processing requests one at a time with immediate responses, batch processing allows you to submit multiple requests together for asynchronous processing. This pattern is particularly useful when:
from line 40
1340
1441The Message Batches API is Anthropic’s first implementation of this pattern.
1542
16<Note>
17 To learn how zero data retention (ZDR) applies to this feature, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
18</Note>
19
2043# Message Batches API
2144
2245The Message Batches API is a powerful, cost-effective way to asynchronously process large volumes of [Messages](https://platform.claude.com/docs/en/api/messages/create) requests. This approach is well-suited to tasks that do not require immediate responses, with most batches finishing in less than 1 hour while reducing costs by 50 percent and increasing throughput.
from line 1839
18161839
18171840* **Result availability:** Batch results are available for 29 days after the batch is created, allowing ample time for retrieval and processing.
18181841
1819## Data retention
1820
1821Batch processing stores request and response data for up to 29 days after batch creation. You can delete a message batch at any time after processing using the [`DELETE /v1/messages/batches/{batch_id}`](https://platform.claude.com/docs/en/api/messages/batches/delete) endpoint. To delete an in-progress batch, cancel it first. Asynchronous processing requires server-side storage of both inputs and outputs until batch completion and result retrieval.
1822
1823For ZDR eligibility across all features, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
1824
18251842## FAQ
18261843
18271844<AccordionGroup>
from line 1883
18661883 Yes, it is possible to use prompt caching with the Message Batches API. However, because asynchronous batch requests can be processed concurrently and in any order, cache hits are provided on a best-effort basis.
18671884 </Accordion>
18681885</AccordionGroup>
1886
1887## Availability
1888
1889Batch processing is available on every model.
1890
1891### API support
1892
1893#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
1894
1895* Availability: Available
1896
1897| Status | Cloud platforms |
1898| --------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
1899| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed) |
1900
1901#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
1902
1903* Availability: Not available
1904
1905### Data handling
1906
1907[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
1908
1909| Item | Zero data retention | HIPAA |
1910| ------- | --------------------------------------------------------------------------------------------------------------------------------------------- | ------------ |
1911| Status | Not eligible | Not eligible |
1912| Details | 29-day retention; async storage required. [Learn more](https://platform.claude.com/docs/en/build-with-claude/batch-processing#data-retention) | — |
1913
1914Applicable endpoints:
1915
1916* `/v1/messages/batches`
1917
1918Batch processing stores request and response data for up to 29 days after batch creation. You can delete a message batch at any time after processing using the [`DELETE /v1/messages/batches/{batch_id}`](https://platform.claude.com/docs/en/api/messages/batches/delete) endpoint. To delete an in-progress batch, cancel it first. Asynchronous processing requires server-side storage of both inputs and outputs until batch completion and result retrieval.
18691919
18701920## Next steps
18711921
build-with-claude/cache-diagnostics Changed · +51 / -4 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling
from line 5
55featureMetadata:
66 status: ga
77 zdr:
8 eligibility: eligible
9 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
8 eligibility: qualified
9 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). Your prompts and Claude's outputs are not stored. A fingerprint of cryptographic hashes and token-count estimates is retained briefly to allow comparison against the next request.
10 supportedModels:
11 - claude-fable-5-1
12 - claude-mythos-5-1
13 - claude-fable-5
14 - claude-mythos-5
15 - claude-mythos-preview
16 - claude-opus-5-5
17 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
21 - claude-haiku-4-5-20251001
22 - claude-opus-4-8
23 - claude-opus-4-7
24 - claude-opus-4-6
25 - claude-sonnet-4-6
26 - claude-opus-4-5-20251101
27 - claude-sonnet-4-5-20250929
1028 supportedPlatforms:
1129 Claude API: ga
1230---
from line 1056
10381056
10391057The API stores a fingerprint only for requests that include the `diagnostics` object. The fingerprint consists only of cryptographic hashes and token-count estimates, keyed by the response `id` and scoped to your organization and workspace. Fingerprints expire after a short period and are not used for any other purpose.
10401058
1041For ZDR eligibility across all features, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
1042
10431059## See also
10441060
10451061* [Prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching)
10461062* [Token counting](https://platform.claude.com/docs/en/build-with-claude/token-counting)
1063
1064## Availability
1065
1066Cache diagnostics are available on every model.
1067
1068### API support
1069
1070#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
1071
1072* Availability: Available
1073
1074| Status | Cloud platforms |
1075| --------- | --------------- |
1076| Available | Claude API |
1077
1078#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
1079
1080* Availability: Not available
1081
1082### Data handling
1083
1084[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
1085
1086| Item | Zero data retention | HIPAA |
1087| ------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------ |
1088| Status | Eligible (qualified; excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Not eligible |
1089| Details | Your prompts and Claude’s outputs are not stored. A fingerprint of cryptographic hashes and token-count estimates is retained briefly to allow comparison against the next request. [Learn more](https://platform.claude.com/docs/en/build-with-claude/cache-diagnostics#data-retention) | — |
1090
1091Applicable endpoints:
1092
1093* `/v1/messages` with `diagnostics`
10471094
build-with-claude/citations Changed · +49 / -0 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling
from line 7
77 zdr:
88 eligibility: eligible
99 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
10 supportedModels:
11 - claude-fable-5-1
12 - claude-mythos-5-1
13 - claude-fable-5
14 - claude-mythos-5
15 - claude-mythos-preview
16 - claude-opus-5-5
17 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
21 - claude-haiku-4-5-20251001
22 - claude-opus-4-8
23 - claude-opus-4-7
24 - claude-opus-4-6
25 - claude-sonnet-4-6
26 - claude-opus-4-5-20251101
27 - claude-sonnet-4-5-20250929
1028 supportedPlatforms:
1129 Claude API: ga
1230 Claude Platform on AWS: ga
1331 Amazon Bedrock: ga
32 Amazon Bedrock (InvokeModel): ga
1433 Claude Models on Google Cloud (Google managed): ga
1534 Microsoft Foundry: ga
1635---
from line 2302
22832302 Upload documents once and reference them by `file_id` across multiple citation requests.
22842303 </Card>
22852304</CardGroup>
2305
2306## Availability
2307
2308Citations are available on every model and eligible for zero data retention and HIPAA readiness.
2309
2310### API support
2311
2312#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
2313
2314* Availability: Available
2315
2316| Status | Cloud platforms |
2317| --------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
2318| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
2319
2320#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
2321
2322* Availability: Not available
2323
2324### Data handling
2325
2326[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
2327
2328| Item | Zero data retention | HIPAA |
2329| ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
2330| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
2331
2332Applicable endpoints:
2333
2334* `/v1/messages`
22862335
build-with-claude/compaction-background Changed · +25 / -3 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) ### Required beta headers
from line 13
1313 - claude-mythos-preview
1414 - claude-opus-5-5
1515 - claude-opus-5
16 - claude-sonnet-5-5
17 - claude-sonnet-5
18 - claude-haiku-5-5
1619 - claude-opus-4-8
1720 - claude-opus-4-7
1821 - claude-opus-4-6
19 - claude-sonnet-5-5
20 - claude-sonnet-5
2122 - claude-sonnet-4-6
22 - claude-haiku-5-5
2323 supportedPlatforms:
2424 Claude API: beta
2525 Claude Platform on AWS: beta
from line 545
545545## Keep thinking valid while the summary is built
546546
547547Turns that arrive while the summary is being written are kept turns. If you send thinking blocks back on a model with [preserved thinking](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking), the thinking in those turns stays valid only while the [conditions for kept thinking](https://platform.claude.com/docs/en/build-with-claude/compaction-thinking-blocks#conditions-for-kept-thinking-to-stay-valid) hold.
548
549## Availability
550
551Background compaction is in beta on select models.
552
553### API support
554
555#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
556
557* Availability: Available
558
559| Status | Cloud platforms | Models |
560| ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
561| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview) |
562
563### Required beta headers
564
565[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
566
567Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
568
569* `compact-2026-09-04`
548570
build-with-claude/compaction-keep-recent-turns Changed · +25 / -3 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) ### Required beta headers
from line 13
1313 - claude-mythos-preview
1414 - claude-opus-5-5
1515 - claude-opus-5
16 - claude-sonnet-5-5
17 - claude-sonnet-5
18 - claude-haiku-5-5
1619 - claude-opus-4-8
1720 - claude-opus-4-7
1821 - claude-opus-4-6
19 - claude-sonnet-5-5
20 - claude-sonnet-5
2122 - claude-sonnet-4-6
22 - claude-haiku-5-5
2323 supportedPlatforms:
2424 Claude API: beta
2525 Claude Platform on AWS: beta
from line 500
500500If you send thinking blocks back on a model with [preserved thinking](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking), the thinking in the kept turns stays valid only while the [conditions for kept thinking](https://platform.claude.com/docs/en/build-with-claude/compaction-thinking-blocks#conditions-for-kept-thinking-to-stay-valid) hold, and one of them limits where the cut can fall.
501501
502502The program’s cut, between a reply and the next `user` message, meets that condition. So does a cut at the end of a request you already made: compact exactly that request’s `messages`, and keep everything your history has gained since.
503
504## Availability
505
506Keep-tail compaction is in beta on select models.
507
508### API support
509
510#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
511
512* Availability: Available
513
514| Status | Cloud platforms | Models |
515| ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
516| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview) |
517
518### Required beta headers
519
520[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
521
522Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
523
524* `compact-2026-09-04`
503525
build-with-claude/compaction-on-demand Changed · +25 / -3 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) ### Required beta headers
from line 13
1313 - claude-mythos-preview
1414 - claude-opus-5-5
1515 - claude-opus-5
16 - claude-sonnet-5-5
17 - claude-sonnet-5
18 - claude-haiku-5-5
1619 - claude-opus-4-8
1720 - claude-opus-4-7
1821 - claude-opus-4-6
19 - claude-sonnet-5-5
20 - claude-sonnet-5
2122 - claude-sonnet-4-6
22 - claude-haiku-5-5
2323 supportedPlatforms:
2424 Claude API: beta
2525 Claude Platform on AWS: beta
from line 828
828828* **Task budgets.** Don’t send the `remaining` value of a [task budget](https://platform.claude.com/docs/en/build-with-claude/task-budgets) (`output_config.task_budget.remaining`) with `compaction` or on requests that carry the block. Doing so returns a 400 error.
829829* **Token counting.** The [token counting](https://platform.claude.com/docs/en/build-with-claude/token-counting) endpoint ignores the `compaction` parameter.
830830* **Content the summary can’t carry.** Images, documents, `container_upload` blocks, and fetched URLs inside the summarized messages are gone once the block replaces them. Restate or re-upload anything a later turn still needs.
831
832## Availability
833
834On-demand compaction is in beta on select models.
835
836### API support
837
838#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
839
840* Availability: Available
841
842| Status | Cloud platforms | Models |
843| ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
844| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview) |
845
846### Required beta headers
847
848[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
849
850Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
851
852* `compact-2026-09-04`
831853
build-with-claude/compaction-thinking-blocks Changed · +25 / -3 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) ### Required beta headers
from line 13
1313 - claude-mythos-preview
1414 - claude-opus-5-5
1515 - claude-opus-5
16 - claude-sonnet-5-5
17 - claude-sonnet-5
18 - claude-haiku-5-5
1619 - claude-opus-4-8
1720 - claude-opus-4-7
1821 - claude-opus-4-6
19 - claude-sonnet-5-5
20 - claude-sonnet-5
2122 - claude-sonnet-4-6
22 - claude-haiku-5-5
2323 supportedPlatforms:
2424 Claude API: beta
2525 Claude Platform on AWS: beta
from line 691
691691```
692692
693693In production, `"drop_block"` keeps requests succeeding when a condition doesn’t hold, and reports each dropped block in `input_transformations` with `reason: "prefix_binding_mismatch"`. An entry whose `path` falls in a kept turn means that turn’s thinking didn’t hold. [What the API does with an invalid block](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#mismatch-behavior) describes what is dropped and says how to alert on it. On Claude Sonnet 5.5 and Claude Haiku 5.5, `block_binding` works only with `thinking: {"type": "adaptive"}`. A request that sets it with `thinking: {"type": "between_tools"}` on Claude Sonnet 5.5, or with `thinking: {"type": "disabled"}` on Claude Haiku 5.5, returns a 400 error. On those requests, make sure every condition holds, or remove the `thinking` and `redacted_thinking` blocks from the kept turns.
694
695## Availability
696
697Compaction and preserved thinking are in beta on select models.
698
699### API support
700
701#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
702
703* Availability: Available
704
705| Status | Cloud platforms | Models |
706| ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
707| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview) |
708
709### Required beta headers
710
711[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
712
713Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
714
715* `compact-2026-09-04`
694716
build-with-claude/compaction-threshold Changed · +57 / -4 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Required beta headers ### Data handling
from line 7
77 betaHeader: compact-2026-01-12
88 zdr:
99 eligibility: eligible
10 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
10 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). Server-side compaction results are returned and round-tripped statelessly through the API response.
1111 supportedModels:
1212 - claude-fable-5-1
1313 - claude-mythos-5-1
from line 16
1616 - claude-mythos-preview
1717 - claude-opus-5-5
1818 - claude-opus-5
19 - claude-sonnet-5-5
20 - claude-sonnet-5
21 - claude-haiku-5-5
1922 - claude-opus-4-8
2023 - claude-opus-4-7
2124 - claude-opus-4-6
22 - claude-sonnet-5-5
23 - claude-sonnet-5
2425 - claude-sonnet-4-6
25 - claude-haiku-5-5
2626 supportedPlatforms:
2727 Claude API: beta
2828 Claude Platform on AWS: beta
2929 Amazon Bedrock: beta
30 Amazon Bedrock (InvokeModel): beta
3031 Claude Models on Google Cloud (Google managed): beta
3132 Microsoft Foundry: beta
3233---
3334
35**API support**: Usage differs by product.
36
37* **Messages API**: Opt in with a `compact_20260112` edit; compaction starts at 150,000 input tokens by default.
38* **Managed Agents**: Always on. It compacts the session’s conversation history to fit the context window.
39
3440Threshold compaction is the automatic kind of compaction: you set a token threshold on your ordinary requests, and the API summarizes older context partway through a request once the threshold is reached. It is supported alongside on-demand compaction, where you decide when the summary is written (see [Compaction on demand](https://platform.claude.com/docs/en/build-with-claude/compaction-on-demand)). To choose between them, see [Choose how to compact](https://platform.claude.com/docs/en/build-with-claude/compaction#choose-how-to-compact).
3541
3642Compaction extends the effective context length for long-running conversations and tasks by automatically summarizing older context when approaching the context window limit. It also keeps the active context small: as a conversation grows, response quality degrades, so compaction replaces older content with a concise summary.
from line 3390
33843390 Explore a practical implementation that manages long-running conversations with instant session memory compaction using background threading and prompt caching.
33853391 </Card>
33863392</CardGroup>
3393
3394## Availability
3395
3396Threshold compaction is in beta on select models and eligible for zero data retention.
3397
3398### API support
3399
3400#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
3401
3402* Availability: Available
3403* Usage: Opt in with a `compact_20260112` edit in `context_management.edits`. Compaction starts when input tokens reach the trigger, 150,000 by default and at least 50,000.
3404
3405| Status | Cloud platforms | Models |
3406| ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
3407| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview) |
3408
3409#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
3410
3411* Availability: Always on
3412* Usage: It compacts the session’s conversation history to fit the context window, and each compaction emits an `agent.thread_context_compacted` event.
3413
3414| Status | Cloud platforms | Models |
3415| ------ | ------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
3416| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview) |
3417
3418### Required beta headers
3419
3420[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
3421
3422Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
3423
3424| Header | APIs |
3425| -------------------- | ------------------------------------------------------------------------------------------- |
3426| `compact-2026-01-12` | [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) |
3427
3428### Data handling
3429
3430[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
3431
3432| Item | Zero data retention | HIPAA |
3433| ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------ |
3434| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Not eligible |
3435| Details | Server-side compaction results are returned and round-tripped statelessly through the API response. | — |
3436
3437Applicable endpoints:
3438
3439* `/v1/messages` with `context_management`
33873440
build-with-claude/context-editing Changed · +66 / -4 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) ### Required beta headers ### Data handling
from line 2
22title: Context editing
33url: https://platform.claude.com/docs/en/build-with-claude/context-editing
44description: Automatically manage conversation context as it grows with context editing.
5featureMetadata:
6 status: beta
7 betaHeader: context-management-2025-06-27
8 zdr:
9 eligibility: eligible
10 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). Context edits (tool use clearing and thinking clearing) are applied in real time.
11 supportedModels:
12 - claude-fable-5-1
13 - claude-mythos-5-1
14 - claude-fable-5
15 - claude-mythos-5
16 - claude-mythos-preview
17 - claude-opus-5-5
18 - claude-opus-5
19 - claude-sonnet-5-5
20 - claude-sonnet-5
21 - claude-haiku-5-5
22 - claude-haiku-4-5-20251001
23 - claude-opus-4-8
24 - claude-opus-4-7
25 - claude-opus-4-6
26 - claude-sonnet-4-6
27 - claude-opus-4-5-20251101
28 - claude-sonnet-4-5-20250929
29 supportedPlatforms:
30 Claude API: beta
31 Claude Platform on AWS: beta
32 Amazon Bedrock: beta
33 Amazon Bedrock (InvokeModel): beta
34 Claude Models on Google Cloud (Google managed): beta
35 Microsoft Foundry: beta
536---
637
7<Note>
8 To learn how zero data retention (ZDR) applies to this feature, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
9</Note>
10
1138## Overview
1239
1340<Note>
from line 2963
29362963 Reduce cost and latency by caching prompt prefixes, and learn how context editing interacts with the cache.
29372964 </Card>
29382965</CardGroup>
2966
2967## Availability
2968
2969Context editing is in beta on every model and eligible for zero data retention.
2970
2971### API support
2972
2973#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
2974
2975* Availability: Available
2976
2977| Status | Cloud platforms |
2978| ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
2979| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
2980
2981### Required beta headers
2982
2983[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
2984
2985Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
2986
2987* `context-management-2025-06-27`
2988
2989### Data handling
2990
2991[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
2992
2993| Item | Zero data retention | HIPAA |
2994| ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------ |
2995| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Not eligible |
2996| Details | Context edits (tool use clearing and thinking clearing) are applied in real time. | — |
2997
2998Applicable endpoints:
2999
3000* `/v1/messages` with `context_management`
29393001
build-with-claude/effort Changed · +104 / -42 lines
## Recommended effort levels ## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling ### Recommended effort levels for Claude Fable 5.1 ### Recommended effort levels for Claude Fable 5 ### Recommended effort levels for Claude Opus 5.5 ### Recommended effort levels for Claude Opus 5 ### Recommended effort levels for Claude Opus 4.8 ### Recommended effort levels for Claude Opus 4.7 ### Recommended effort levels for Claude Sonnet 5.5 ### Recommended effort levels for Claude Sonnet 5 ### Recommended effort levels for Claude Sonnet 4.6 ### Recommended effort levels for Claude Haiku 5.5
from line 15
1515 - claude-mythos-preview
1616 - claude-opus-5-5
1717 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
1821 - claude-opus-4-8
1922 - claude-opus-4-7
2023 - claude-opus-4-6
21 - claude-opus-4-5-20251101
22 - claude-sonnet-5-5
23 - claude-sonnet-5
2424 - claude-sonnet-4-6
25 - claude-haiku-5-5
25 - claude-opus-4-5-20251101
2626 supportedPlatforms:
2727 Claude API: ga
2828 Claude Platform on AWS: ga
2929 Amazon Bedrock: ga
30 Amazon Bedrock (InvokeModel): ga
3031 Claude Models on Google Cloud (Google managed): ga
3132 Microsoft Foundry: ga
3233---
3334
35**API support**: Usage differs by product.
36
37* **Messages API**: Always on. Set `output_config.effort` on the request. With no value, the model’s default level applies.
38* **Managed Agents**: Always on. Set `effort` on the agent’s model object, or in a session’s model override. With no value, Managed Agents fills in the model’s default.
39
3440The effort parameter lets you control how many tokens Claude spends when responding to requests. You can trade off between response thoroughness and token efficiency with a single model. The top-level effort parameter is available on all supported models with no beta header required. [Per-message effort](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) is in beta.
3541
3642<Tip>
from line 256
250256
251257The per-model recommendations that follow override this table where they differ.
252258
253### Recommended effort levels for Claude Fable 5.1
259## Recommended effort levels
254260
255Claude Fable 5.1 supports all five effort levels. **Start with `high`, the default.** Step up to `xhigh` or `max` for the most capability-sensitive agentic and coding work, and step down to `medium` or `low` for routine or latency-sensitive work once your evals show quality holds. At `high` and above, set a large `max_tokens`. It’s a hard limit on total output (thinking plus response text). The same recommendations apply to Claude Mythos 5.1. See [Prompting Claude Fable 5.1](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5-1#consider-all-effort-levels).
261**[Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) & [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview)** — *For demanding reasoning and long-horizon agentic work*
256262
263Claude Fable 5.1 supports all five effort levels. **Start with `high`, the default.** Step up to `xhigh` or `max` for the most capability-sensitive agentic and coding work, and step down to `medium` or `low` for routine or latency-sensitive work once your evals show quality holds. At `high` and above, set a large `max_tokens`. It’s a hard limit on total output (thinking plus response text).
264
257265Claude Fable 5.1 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache.
258266
259### Recommended effort levels for Claude Fable 5
267See also: [Prompting Claude Fable 5.1](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5-1#consider-all-effort-levels)
260268
261Effort is the primary control for trading off intelligence, latency, and cost on Claude Fable 5. **Start with `high`, the default, for most tasks**, use `xhigh` for the most capability-sensitive workloads, and step down to `medium` or `low` for routine work. Lower effort settings on Claude Fable 5 still perform well and often exceed `xhigh` performance on prior models. At `high` and `xhigh`, set a large `max_tokens`. It’s a hard limit on total output (thinking plus response text). See [Cost control](https://platform.claude.com/docs/en/build-with-claude/thinking-steering-and-cost#cost-control).
269**[Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview)** — *For long-running agentic coding and knowledge work*
262270
263Reduce effort if a task completes but takes longer than necessary, or if you want a faster, more interactive working style. The same recommendations apply to Claude Mythos 5. For fuller guidance, see [Prompting Claude Fable 5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5).
271Claude Opus 5.5 supports all five effort levels, and `medium` is the default (Claude Opus 5 and earlier Opus models default to `high`, so a request that omits `effort` runs one level lower than it did on Claude Opus 5). Adaptive thinking is always on and can’t be turned off, so effort is the primary control for how much the model reasons and what a request costs. Run an effort sweep on your own evals rather than carrying settings over from an earlier model, and set a large `max_tokens` at the higher levels: it’s a hard limit on total output (thinking plus response text). Requests that set `thinking: {"type": "disabled"}` return a 400 error at every effort level. Claude Opus 5.5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache.
264272
265### Recommended effort levels for Claude Opus 5.5
273See also: [Prompting Claude Opus 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5-5)
266274
267Claude Opus 5.5 supports all five effort levels, and `medium` is the default (Claude Opus 5 and earlier Opus models default to `high`, so a request that omits `effort` runs one level lower than it did on Claude Opus 5). Adaptive thinking is always on and can’t be turned off, so effort is the primary control for how much the model reasons and what a request costs. Run an effort sweep on your own evals rather than carrying settings over from an earlier model, and set a large `max_tokens` at the higher levels: it’s a hard limit on total output (thinking plus response text). Requests that set `thinking: {"type": "disabled"}` return a 400 error at every effort level. Claude Opus 5.5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache. See [Prompting Claude Opus 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5-5).
275**[Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview)** — *The best combination of speed and intelligence*
268276
269### Recommended effort levels for Claude Opus 5
277Claude Sonnet 5.5 supports all five effort levels, and `high` is the default on the Claude API. Its levels are recalibrated, so a level doesn’t produce the same amount of thinking as the same level on Claude Sonnet 5. Run a fresh effort sweep on your evals rather than carrying over the setting you used on Claude Sonnet 5. Start with `high` unless your workload is agentic or latency-sensitive. For agentic coding and multistep tool use, start with `medium` for well-specified tasks and move to `high` for harder or longer ones. For chat and other latency-sensitive work, start with `medium` or `low`. Use `xhigh` or `max` only where your evals show a quality gain. Set `max_tokens` with room for thinking and the reply. Thinking counts toward `max_tokens` even when the thinking content isn’t returned. For agentic coding, set `max_tokens` to 128,000, the model’s maximum, and [stream](https://platform.claude.com/docs/en/build-with-claude/streaming) the response.
270278
279To turn off up-front thinking, send `thinking: {"type": "between_tools"}` instead of `"disabled"`. It’s the lowest thinking setting on Claude Sonnet 5.5, and it works at `low`, `medium`, and `high` effort. At `xhigh` or `max`, a request with it returns a 400 error, so use adaptive thinking at those levels: omit the `thinking` field or send `thinking: {"type": "adaptive"}`. See [Running without up-front thinking](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-sonnet-5-5#running-without-up-front-thinking).
280
281Claude Sonnet 5.5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache. With `between_tools`, effort can’t change mid-conversation: a per-message `output_config.effort` that differs from the level in effect returns a 400 error. To vary effort for each turn, use adaptive thinking.
282
283See also: [Prompting Claude Sonnet 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-sonnet-5-5#calibrate-effort)
284
285**[Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview)** — *For high-volume, latency-sensitive tasks such as classification, extraction, and routing*
286
287Claude Haiku 5.5 supports all five effort levels, and `medium` is the default. Effort is the main control for how much the model thinks, and with it quality, latency, and cost. **Start with `medium`** for most work, including agentic coding. Use `low`, the cheapest and fastest level, for chat, short tool tasks, and simple, high-volume requests. In long agent prompts, the model is more likely to skip a search, stop early, or skip a check at `low`. Use `high` for knowledge work, longer agent tasks, and strict instruction following. Use `xhigh` or `max` only where your evals show a quality gain, and compare them with Claude Sonnet 5.5 on performance, cost, and speed. Thinking is on by default and counts toward `max_tokens`, so leave room for it.
288
289To get less thinking, lower the effort level. You can also send `thinking: {"type": "disabled"}` at `high` effort or below. At `xhigh` or `max`, it returns a 400 error, so use adaptive thinking there: omit the `thinking` field or send `thinking: {"type": "adaptive"}`.
290
291On the Claude API and Claude Models on Google Cloud (Google managed), Claude Haiku 5.5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache. With `thinking: {"type": "disabled"}`, effort can’t change mid-conversation: a per-message `output_config.effort` that differs from the level in effect returns a 400 error. To vary effort for each turn, use adaptive thinking.
292
293See also: [Prompting Claude Haiku 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-haiku-5-5#use-effort-to-control-thinking)
294
295**[Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview) & [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview)**
296
297Effort is the primary control for trading off intelligence, latency, and cost on Claude Fable 5. **Start with `high`, the default, for most tasks**, use `xhigh` for the most capability-sensitive workloads, and step down to `medium` or `low` for routine work. Lower effort settings on Claude Fable 5 still perform well and often exceed `xhigh` performance on prior models. At `high` and `xhigh`, set a large `max_tokens`. It’s a hard limit on total output (thinking plus response text). See [Cost control](https://platform.claude.com/docs/en/build-with-claude/thinking-steering-and-cost#cost-control).
298
299Reduce effort if a task completes but takes longer than necessary, or if you want a faster, more interactive working style.
300
301See also: [Prompting Claude Fable 5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5)
302
303**[Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview)**
304
271305Claude Opus 5 supports all five effort levels. **Start with `high`, the default**, and adjust based on your evals: step up to `xhigh` for demanding coding and agentic work, or to `max` when a task justifies unconstrained token spending. Use `low` and `medium` liberally as your primary control for token cost and response time wherever your evals show quality holds. If you carried effort settings over from an earlier model, run a fresh effort sweep on your evals rather than reusing them.
272306
273307Effort controls thinking volume, not visible response length: on Claude Opus 5, changing effort does not reliably shorten responses, so [prompt for length](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5#response-length-and-verbosity) instead.
from line 314
280314
281315Claude Opus 5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache. Per-message effort isn’t available for Claude Opus 5 on Amazon Bedrock.
282316
283### Recommended effort levels for Claude Opus 4.8
317See also: [Prompting Claude Opus 5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5)
284318
319**[Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview)**
320
285321The guidance for Claude Opus 4.7 also applies to Claude Opus 4.8. **Start with `xhigh` for coding and agentic use cases**, use `high` for most other intelligence-sensitive workloads, and step down to `medium` or `low` only when you’ve measured that the lower level holds quality on your evals.
286322
287323The API default is `high`. Set `effort` explicitly to use a different level. The value you pass overrides the default.
from line 324
288324
289325When running Claude Opus 4.8 at `xhigh` or `max` effort, set a large `max_tokens` so the model has room to think and act across subagents and tool calls. Starting at 64k tokens and tuning from there is a reasonable default.
290326
291### Recommended effort levels for Claude Opus 4.7
327See also: [Prompting Claude Opus 4.8](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-4-8)
292328
329**[Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview)**
330
293331**Start with `xhigh` for coding and agentic use cases**, and use `high` as the minimum for most intelligence-sensitive workloads. Step down to `medium` for cost-sensitive workloads, or up to `max` only when your evals show measurable headroom at `xhigh`.
294332
295333The API default is `high`. To use `xhigh`, set `effort` explicitly. The value you pass overrides the default.
296334
297| Effort | Guidance for Claude Opus 4.7 |
298| -------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
299| `low` | Efficient, but best for short, scoped tasks. Pair `low` with explicit checklists if your task has multiple sections. |
300| `medium` | The setting for the average workflow where you want good results while reducing costs. |
301| `high` | Advanced use cases that still need a balance of intelligence and token consumption. This is often the best balance of quality and token efficiency. |
302| `xhigh` | The recommended starting point for coding and agentic work, and for exploratory tasks such as repeated tool calling, detailed web search, and knowledge-base search. Expect meaningfully higher token usage than `high`. |
303| `max` | Reserve for frontier problems. On most workloads `max` adds significant cost for relatively small quality gains, and on some structured-output or less intelligence-sensitive tasks it can lead to overthinking. |
335* **Low effort:** Efficient, but best for short, scoped tasks. Pair `low` with explicit checklists if your task has multiple sections.
336* **Medium effort:** The setting for the average workflow where you want good results while reducing costs.
337* **High effort:** Advanced use cases that still need a balance of intelligence and token consumption. This is often the best balance of quality and token efficiency.
338* **Xhigh effort:** The recommended starting point for coding and agentic work, and for exploratory tasks such as repeated tool calling, detailed web search, and knowledge-base search. Expect meaningfully higher token usage than `high`.
339* **Max effort:** Reserve for frontier problems. On most workloads `max` adds significant cost for relatively small quality gains, and on some structured-output or less intelligence-sensitive tasks it can lead to overthinking.
304340
305341Claude Opus 4.7 also respects effort levels more strictly than Claude Opus 4.6, especially at `low` and `medium`. At lower effort levels, the model scopes its work to what was asked rather than doing more than requested. If you observe shallow reasoning on complex problems with Claude Opus 4.7, raise effort rather than prompting around it. If you must keep effort low for latency, add targeted guidance such as “This task involves multistep reasoning. Think carefully before responding.”
306342
307343When running Claude Opus 4.7 at `xhigh` or `max` effort, set a large `max_tokens` so the model has room to think and act across subagents and tool calls. Starting at 64k tokens and tuning from there is a reasonable default.
308344
309### Recommended effort levels for Claude Sonnet 5.5
345See also: [Prompting best practices](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/claude-prompting-best-practices)
310346
311Claude Sonnet 5.5 supports all five effort levels, and `high` is the default on the Claude API. Its levels are recalibrated, so a level doesn’t produce the same amount of thinking as the same level on Claude Sonnet 5. Run a fresh effort sweep on your evals rather than carrying over the setting you used on Claude Sonnet 5. Start with `high` unless your workload is agentic or latency-sensitive. For agentic coding and multistep tool use, start with `medium` for well-specified tasks and move to `high` for harder or longer ones. For chat and other latency-sensitive work, start with `medium` or `low`. Use `xhigh` or `max` only where your evals show a quality gain. Set `max_tokens` with room for thinking and the reply. Thinking counts toward `max_tokens` even when the thinking content isn’t returned. For agentic coding, set `max_tokens` to 128,000, the model’s maximum, and [stream](https://platform.claude.com/docs/en/build-with-claude/streaming) the response.
347**[Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview)**
312348
313To turn off up-front thinking, send `thinking: {"type": "between_tools"}` instead of `"disabled"`. It’s the lowest thinking setting on Claude Sonnet 5.5, and it works at `low`, `medium`, and `high` effort. At `xhigh` or `max`, a request with it returns a 400 error, so use adaptive thinking at those levels: omit the `thinking` field or send `thinking: {"type": "adaptive"}`. See [Running without up-front thinking](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-sonnet-5-5#running-without-up-front-thinking).
314
315Claude Sonnet 5.5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache. With `between_tools`, effort can’t change mid-conversation: a per-message `output_config.effort` that differs from the level in effect returns a 400 error. To vary effort for each turn, use adaptive thinking. See [Prompting Claude Sonnet 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-sonnet-5-5#calibrate-effort).
316
317### Recommended effort levels for Claude Sonnet 5
318
319349Claude Sonnet 5 defaults to `high` effort on the Claude API and Claude Code.
320350
321351* **High effort (default):** Suitable for complex reasoning, coding, and agentic tasks where quality matters more than speed or cost.
322* **Xhigh effort:** For the hardest coding and agentic tasks. See [Prompting Claude Sonnet 5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-sonnet-5#calibrating-effort-and-thinking-depth).
352* **Xhigh effort:** For the hardest coding and agentic tasks.
323353* **Medium effort:** Cost-saving step-down from the default. Comparable to Claude Sonnet 4.6 at high effort.
324354* **Low effort:** For high-volume or latency-sensitive workloads. Suitable for chat and non-coding use cases where faster turnaround is prioritized.
325355* **Max effort:** For tasks requiring the absolute highest capability with no constraints on token spending.
326356
327### Recommended effort levels for Claude Sonnet 4.6
357See also: [Prompting Claude Sonnet 5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-sonnet-5#calibrating-effort-and-thinking-depth)
328358
359**[Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview)**
360
329361Sonnet 4.6 defaults to `high` effort. Explicitly set effort when using Sonnet 4.6 to avoid unexpected latency:
330362
331363* **Medium effort (recommended default):** Best balance of speed, cost, and performance for most applications. Suitable for agentic coding, tool-heavy workflows, and code generation.
from line 365
333365* **High effort:** For complex reasoning and tasks where quality matters more than speed or cost.
334366* **Max effort:** For tasks requiring the absolute highest capability with no constraints on token spending.
335367
336### Recommended effort levels for Claude Haiku 5.5
368See also: [Prompting best practices](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/claude-prompting-best-practices)
337369
338Claude Haiku 5.5 supports all five effort levels, and `medium` is the default. Effort is the main control for how much the model thinks, and with it quality, latency, and cost. **Start with `medium`** for most work, including agentic coding. Use `low`, the cheapest and fastest level, for chat, short tool tasks, and simple, high-volume requests. In long agent prompts, the model is more likely to skip a search, stop early, or skip a check at `low`. Use `high` for knowledge work, longer agent tasks, and strict instruction following. Use `xhigh` or `max` only where your evals show a quality gain, and compare them with Claude Sonnet 5.5 on performance, cost, and speed. Thinking is on by default and counts toward `max_tokens`, so leave room for it. See [Prompting Claude Haiku 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-haiku-5-5#use-effort-to-control-thinking).
339
340To get less thinking, lower the effort level. You can also send `thinking: {"type": "disabled"}` at `high` effort or below. At `xhigh` or `max`, it returns a 400 error, so use adaptive thinking there: omit the `thinking` field or send `thinking: {"type": "adaptive"}`.
341
342On the Claude API and Claude Models on Google Cloud (Google managed), Claude Haiku 5.5 also supports [changing effort mid-conversation](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) with a per-message `output_config`, which preserves the prompt cache. With `thinking: {"type": "disabled"}`, effort can’t change mid-conversation: a per-message `output_config.effort` that differs from the level in effect returns a 400 error. To vary effort for each turn, use adaptive thinking.
343
344370## Effort with tool use
345371
346372When using tools, the effort parameter affects both the explanations around tool calls and the tool calls themselves. Lower effort levels tend to:
from line 704
678704 Understand how thinking works, when Claude thinks by default, and how thinking interacts with effort.
679705 </Card>
680706</CardGroup>
707
708## Availability
709
710Effort is available on select models and eligible for zero data retention and HIPAA readiness.
711
712### API support
713
714#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
715
716* Availability: Always on
717* Usage: Set `output_config.effort` on the request. With no value, the model’s default level applies.
718
719| Status | Cloud platforms | Models |
720| --------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
721| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview), [Claude Opus 4.5](https://platform.claude.com/docs/en/models/opus-4-5/overview) |
722
723#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
724
725* Availability: Always on
726* Usage: Set `effort` on the agent’s model object, or in a session’s model override. With no value, Managed Agents fills in the model’s default.
727
728| Status | Cloud platforms | Models |
729| --------- | ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
730| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview), [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview), [Claude Opus 4.5](https://platform.claude.com/docs/en/models/opus-4-5/overview) |
731
732### Data handling
733
734[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
735
736| Item | Zero data retention | HIPAA |
737| ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
738| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
739
740Applicable endpoints:
741
742* `/v1/messages` with `effort`
681743
build-with-claude/extended-thinking Changed · +42 / -11 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
from line 2
22title: Extended thinking
33url: https://platform.claude.com/docs/en/build-with-claude/extended-thinking
44description: Configure manual extended thinking with a fixed budget_tokens budget on Claude models that support it, and migrate to adaptive thinking.
5featureMetadata:
6 supportedModels:
7 - claude-mythos-preview
8 - claude-haiku-4-5-20251001
9 - claude-opus-4-6
10 - claude-sonnet-4-6
11 - claude-opus-4-5-20251101
12 - claude-sonnet-4-5-20250929
13 supportedPlatforms:
14 Claude API: ga
15 Claude Platform on AWS: ga
16 Claude Models on Google Cloud (Anthropic managed): ga
17 Amazon Bedrock: ga
18 Amazon Bedrock (InvokeModel): ga
19 Claude Models on Google Cloud (Google managed): ga
20 Microsoft Foundry: ga
521---
622
7<Note>
8 To learn how zero data retention (ZDR) applies to this feature, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
9</Note>
10
1123<Warning>
12 Extended thinking (`thinking.type: "enabled"` with `budget_tokens`) is deprecated on the Claude 4.6 models (requests using it still succeed). Claude 4.7 and later models do not support it and reject requests that use it, returning a 400 error. On Claude 4.5 and earlier models that support thinking, extended thinking is the only available thinking mode. Claude Mythos Preview supports both modes. Where both modes are available, use [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) instead.
24 Extended thinking is a legacy feature.
1325
14 See [Migrating to adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking#migrating-to-adaptive-thinking) to move to adaptive thinking. If your model supports only extended thinking, this page describes the supported configuration; no change is needed until you move to a newer model.
26 Extended thinking (`thinking.type: "enabled"` with `budget_tokens`) is deprecated. Claude 4.7 and later models do not support it and reject requests that use it, returning a 400 error.
27
28 * Recommended alternative: [Adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking-steering-and-cost)
29 * Migration guide: [Migrating to adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking#migrating-to-adaptive-thinking)
1530</Warning>
1631
17<Note>
18 If a request fails with a 400 error whose message starts with `"thinking.type.enabled" is not supported`, your model uses adaptive thinking instead. See [Troubleshooting thinking](https://platform.claude.com/docs/en/build-with-claude/thinking-troubleshooting#error-thinking-type-enabled), or jump to [Migrating to adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking#migrating-to-adaptive-thinking).
19</Note>
20
2132Extended thinking in manual mode gives you direct control over how much Claude thinks. You set a thinking token budget on each request with `thinking: {type: "enabled", budget_tokens: N}`, and Claude thinks against that budget before it starts its final answer. Manual mode remains useful when your workload requires predictable latency or precise control over thinking costs. This page covers how to set and tune the budget, how manual mode interacts with interleaved thinking and prompt caching, and how to migrate to adaptive thinking.
2233
2334To learn how thinking itself works, including thinking blocks and the response shape, the `display` parameter, streaming, thinking with tool use, and encryption, see the [thinking overview](https://platform.claude.com/docs/en/build-with-claude/thinking).
from line 416
405416 Preserve thinking blocks and manage thinking across tool calls and turns.
406417 </Card>
407418</CardGroup>
419
420## Availability
421
422Extended thinking is available or deprecated on select models.
423
424### API support
425
426#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
427
428* Availability: Available
429
430| Status | Models |
431| ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
432| Available | Claude Mythos Preview, [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview), [Claude Opus 4.5](https://platform.claude.com/docs/en/models/opus-4-5/overview), [Claude Sonnet 4.5](https://platform.claude.com/docs/en/models/sonnet-4-5/overview) |
433| Deprecated | [Claude Opus 4.6](https://platform.claude.com/docs/en/models/opus-4-6/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview) |
434
435#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
436
437* Availability: Not available
438* Usage: On Claude 4.5 models, Managed Agents turns on manual thinking itself with a fixed budget you can’t change. On newer models it uses adaptive thinking. Use [Effort](https://platform.claude.com/docs/en/build-with-claude/effort) instead.
408439
build-with-claude/fallback-credit Changed · +53 / -0 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Required beta headers
from line 2
22title: Fallback credit
33url: https://platform.claude.com/docs/en/build-with-claude/fallback-credit
44description: Avoid paying the prompt-cache cost twice when you retry a refused request on another model.
5featureMetadata:
6 status: beta
7 betaHeader: fallback-credit-2026-07-01, server-side-fallback-2026-07-01, fallback-credit-2026-06-01
8 supportedModels:
9 - claude-fable-5-1
10 - claude-mythos-5-1
11 - claude-fable-5
12 - claude-opus-5-5
13 - claude-opus-5
14 - claude-sonnet-5-5
15 supportedPlatforms:
16 Claude API: beta
17 Claude Platform on AWS: beta
18 Amazon Bedrock:
19 availability: beta
20 note: On Amazon Bedrock and Google Cloud the token is bound to your caller identity (there are no workspaces), and `allowed_fallback_models` is not exposed.
21 Claude Models on Google Cloud (Google managed):
22 availability: beta
23 note: On Amazon Bedrock and Google Cloud the token is bound to your caller identity (there are no workspaces), and `allowed_fallback_models` is not exposed.
24 Microsoft Foundry: beta
525---
626
727Prompt caches are per-model. When a model declines a request and you retry on another model, the conversation prefix already cached for the first model must be written into the new model’s cache in full. Cache writes cost more than cache reads. Fallback credit removes that extra cost. The refusal carries a credit token, you echo the token on the retry, and the retry is billed as though the conversation had been on the new model from the start.
from line 641
621641 The SDK helper that applies fallback credit automatically.
622642 </Card>
623643</CardGroup>
644
645## Availability
646
647Fallback credit is in beta on select models.
648
649### API support
650
651#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
652
653* Availability: Available
654* Usage: Needed only when you write the retry yourself: send the `fallback-credit-2026-07-01` beta header and put the refusal’s `fallback_credit_token` on the retry.
655
656| Status | Cloud platforms | Models |
657| ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
658| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview) |
659| Beta<sup>1</sup> | [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview) |
660
6611. On Amazon Bedrock and Google Cloud the token is bound to your caller identity (there are no workspaces), and `allowed_fallback_models` is not exposed.
662
663#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
664
665* Availability: Not available
666
667### Required beta headers
668
669[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
670
671Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
672
673| Header | Cloud platforms |
674| ------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
675| `fallback-credit-2026-07-01` or `server-side-fallback-2026-07-01` or `fallback-credit-2026-06-01` | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
676| `fallback-credit-2026-07-01` or `fallback-credit-2026-06-01` | [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed) |
624677
build-with-claude/fast-mode Changed · +81 / -45 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Required beta headers ### Data handling ## Supported platforms ## Model availability ### Claude Opus 5.5 ### Claude Opus 5 and Claude Opus 4.8
from line 3
33url: https://platform.claude.com/docs/en/build-with-claude/fast-mode
44description: Get up to 2.5x higher output tokens per second from supported Claude models.
55featureMetadata:
6 betaHeader: fast-mode-2026-02-01
67 zdr:
78 eligibility: eligible
8 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
9 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). Same Messages API endpoint with faster inference. ZDR applies regardless of speed setting.
910 supportedModels:
1011 - claude-opus-5-5
1112 - claude-opus-5
1213 - claude-opus-4-8
1314 supportedPlatforms:
14 Claude API:
15 availability: ga
16 note: Includes [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview).
17 Microsoft Foundry:
18 availability: ga
19 note: Available for Claude Opus 5.5 on deployments hosted on Azure, not on deployments [hosted on Anthropic](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry#hosting-options).
15 Claude API: ga
16 Claude Models on Google Cloud (Anthropic managed): ga
17 Microsoft Foundry (hosted on Azure): ga
2018 details:
2119 - Fast mode on Claude Opus 5 and Claude Opus 4.8 is in research preview and needs the `fast-mode-2026-02-01` beta header.
2220---
2321
22**API support**: Usage differs by product.
23
24* **Messages API**: Set `speed: "fast"` on the request.
25* **Managed Agents**: Set `speed` on the agent’s model object, for example `{"id": "claude-opus-5-5", "speed": "fast"}`.
26
2427Fast mode delivers up to 2.5x higher output tokens per second at premium pricing. Set `speed: "fast"` on your request to opt in. Fast mode is available on Claude Opus 5.5 and in research preview on Claude Opus 5 and Claude Opus 4.8.
2528
2629## How fast mode works
from line 36
3336
3437## Basic usage
3538
36Fast mode works only on supported models. See [Model availability](https://platform.claude.com/docs/en/build-with-claude/fast-mode#supported-models) for details, including which models need a beta header.
39Fast mode works only on supported models. See [Availability](https://platform.claude.com/docs/en/build-with-claude/fast-mode#feature-availability) for details, including which models need a beta header.
3740
3841<CodeGroup>
3942 ```bash cURL
from line 196
193196 ```
194197</CodeGroup>
195198
196## Supported platforms
197
198Fast mode is available on the Claude API, including [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview).
199
200Fast mode on Claude Opus 5.5 is also available on [Claude in Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) for deployments hosted on Azure. It is not available for deployments [hosted on Anthropic](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry#hosting-options).
201
202Fast mode on Claude Opus 5.5 is also available on [Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed).
203
204## Model availability
205
206Fast mode is available only on the following models. If you were using fast mode on an earlier model, [migrate to a model that supports fast mode](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide).
207
208### Claude Opus 5.5
209
210Fast mode is available on Claude Opus 5.5. It doesn’t need a beta header or an access request.
211
212### Claude Opus 5 and Claude Opus 4.8
213
214Fast mode on Claude Opus 5 and Claude Opus 4.8 is in research preview. Requests that set `speed` must send the `fast-mode-2026-02-01` [beta header](https://platform.claude.com/docs/en/api/beta-headers); without it, they return a `400` error. To access fast mode on Claude Opus 5 or Claude Opus 4.8, contact your account manager or [join the waitlist](https://claude.com/fast-mode).
215
216## Pricing
217
218Fast mode is priced at a multiplier on standard rates across the full context window, including requests over 200k input tokens.
219
220| Model | Input | Output |
221| ------------------------------- | -------------- | -------------- |
222| Claude Opus 5.5 | $8 USD / MTok | $40 USD / MTok |
223| Claude Opus 5 / Claude Opus 4.8 | $10 USD / MTok | $50 USD / MTok |
224
225Fast mode pricing stacks with other pricing modifiers:
226
227* [Prompt caching multipliers](https://platform.claude.com/docs/en/about-claude/pricing#prompt-caching) apply on top of fast mode pricing.
228* [Data residency](https://platform.claude.com/docs/en/manage-claude/data-residency) multipliers apply on top of fast mode pricing.
229
230For complete pricing details, see the [Pricing](https://platform.claude.com/docs/en/about-claude/pricing#fast-mode-pricing) page.
231
232199## Rate limits
233200
234201Fast mode has a dedicated rate limit that is separate from standard Opus rate limits. When your fast mode rate limit is exceeded, the API returns a `429` error with a `retry-after` header indicating when capacity will be available.
from line 217
250217
251218## Checking which speed was used
252219
253The response `usage` object includes a `speed` field that indicates which speed was used, either `"fast"` or `"standard"`. Requesting `speed: "fast"` on a [model that doesn’t support fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode#supported-models) returns an error, and so does exceeding fast mode’s rate limits or capacity (a `429` or `529`). When a request with `speed: "fast"` succeeds, `usage.speed` is `"fast"`. Claude Opus 4.6 is an exception: instead of returning an error, a request with `speed: "fast"` and the `fast-mode-2026-02-01` beta header succeeds, but runs at standard speed. This behavior is correctly reflected in `usage.speed`, which shows `"standard"` in this situation. On Claude Models on Google Cloud (Anthropic managed), a Claude Opus 4.6 request with `speed: "fast"` returns a `400` error even with the header.
220The response `usage` object includes a `speed` field that indicates which speed was used, either `"fast"` or `"standard"`. Requesting `speed: "fast"` on a [model that doesn’t support fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode#availability-feature-support) returns an error, and so does exceeding fast mode’s rate limits or capacity (a `429` or `529`). When a request with `speed: "fast"` succeeds, `usage.speed` is `"fast"`. Claude Opus 4.6 is an exception: instead of returning an error, a request with `speed: "fast"` and the `fast-mode-2026-02-01` beta header succeeds, but runs at standard speed. This behavior is correctly reflected in `usage.speed`, which shows `"standard"` in this situation. On Claude Models on Google Cloud (Anthropic managed), a Claude Opus 4.6 request with `speed: "fast"` returns a `400` error even with the header.
254221
255222<CodeGroup>
256223 ```bash cURL
from line 663
696663
697664* **Batch API:** Fast mode is not available with the [Batch API](https://platform.claude.com/docs/en/build-with-claude/batch-processing).
698665* **Priority Tier:** Fast mode is not available with a [Priority Tier](https://platform.claude.com/docs/en/api/service-tiers) commitment.
666
667## Pricing
668
669Fast mode is priced at a multiplier on standard rates across the full context window, including requests over 200k input tokens.
670
671| Model | Input | Output |
672| ------------------------------- | -------------- | -------------- |
673| Claude Opus 5.5 | $8 USD / MTok | $40 USD / MTok |
674| Claude Opus 5 / Claude Opus 4.8 | $10 USD / MTok | $50 USD / MTok |
675
676Fast mode pricing stacks with other pricing modifiers:
677
678* [Prompt caching multipliers](https://platform.claude.com/docs/en/about-claude/pricing#prompt-caching) apply on top of fast mode pricing.
679* [Data residency](https://platform.claude.com/docs/en/manage-claude/data-residency) multipliers apply on top of fast mode pricing.
680
681For complete pricing details, see the [Pricing](https://platform.claude.com/docs/en/about-claude/pricing#fast-mode-pricing) page.
682
683## Availability
684
685Fast mode is available or in research preview on select models and eligible for zero data retention and HIPAA readiness.
686
687### API support
688
689Fast mode is available only on the following models. If you were using fast mode on an earlier model, [migrate to a model that supports fast mode](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide).
690
691To access fast mode on Claude Opus 5 or Claude Opus 4.8, contact your account manager or [join the waitlist](https://claude.com/fast-mode).
692
693#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
694
695* Availability: Available
696* Usage: Set `speed: "fast"` on the request.
697
698| Status | Cloud platforms | Models |
699| ---------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
700| Available | Claude API, [Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed), [Microsoft Foundry (hosted on Azure)](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) |
701| Research preview | Claude API | [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview) |
702
703#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
704
705* Availability: Available
706* Usage: Set `speed` on the agent’s model object, for example `{"id": "claude-opus-5-5", "speed": "fast"}`.
707
708| Status | Cloud platforms | Models |
709| ---------------- | --------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
710| Available | Claude API | [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) |
711| Research preview | Claude API | [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview) |
712
713### Required beta headers
714
715[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
716
717Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
718
719| Header | Models |
720| ---------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------ |
721| `fast-mode-2026-02-01` | [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview) |
722
723### Data handling
724
725[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
726
727| Item | Zero data retention | HIPAA |
728| ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
729| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
730| Details | Same Messages API endpoint with faster inference. ZDR applies regardless of speed setting. | — |
731
732Applicable endpoints:
733
734* `/v1/messages` with `speed: "fast"`
699735
700736## Next steps
701737
build-with-claude/files Changed · +69 / -8 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling
from line 4
44description: Upload files once, reference them by file_id in Messages requests, and download outputs created by skills or the code execution tool.
55featureMetadata:
66 status: ga
7 zdr: not-eligible
7 zdr:
8 eligibility: not-eligible
9 note: Files retained until explicitly deleted or they reach their configured expiration.
10 supportedModels:
11 - claude-fable-5-1
12 - claude-mythos-5-1
13 - claude-fable-5
14 - claude-mythos-5
15 - claude-mythos-preview
16 - claude-opus-5-5
17 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
21 - claude-haiku-4-5-20251001
22 - claude-opus-4-8
23 - claude-opus-4-7
24 - claude-opus-4-6
25 - claude-sonnet-4-6
26 - claude-opus-4-5-20251101
27 - claude-sonnet-4-5-20250929
828 supportedPlatforms:
929 Claude API: ga
1030 Claude Platform on AWS: ga
11 Microsoft Foundry:
12 availability: ga
13 note: On [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry), the Files API requires a [Hosted on Anthropic deployment](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry#additional-features-not-supported-when-hosted-on-azure).
31 Claude Models on Google Cloud (Anthropic managed): ga
32 Microsoft Foundry (hosted on Anthropic): ga
1433---
1534
35**API support**: Usage differs by product.
36
37* **Messages API**: Refer to an uploaded file by its `file_id` in a `document`, `image` or `container_upload` block.
38* **Managed Agents**: Mount files in the session’s sandbox, or refer to a `file_id` in a `user.message` event.
39
1640The Files API lets you upload and manage files to use with the Claude API without re-uploading content with each request. This is particularly useful when using the [code execution tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/code-execution-tool) to provide inputs (for example, datasets and documents) and then download outputs (for example, charts). You can [explore the API reference directly](https://platform.claude.com/docs/en/api/files/upload), in addition to this guide.
1741
1842## File type support
1943
20Referencing a `file_id` in a Messages request is supported on all models that support the given file type. [Images](https://platform.claude.com/docs/en/build-with-claude/vision) are supported on all current Claude models. For [PDFs](https://platform.claude.com/docs/en/build-with-claude/pdf-support) and [other file types with the code execution tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/code-execution-tool#compatibility), see the linked pages for model support.
44Referencing a `file_id` in a Messages request is supported on all models that support the given file type. [Images](https://platform.claude.com/docs/en/build-with-claude/vision) are supported on all current Claude models. For [PDFs](https://platform.claude.com/docs/en/build-with-claude/pdf-support) and [other file types with the code execution tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/code-execution-tool#availability-feature-support), see the linked pages for model support.
2145
2246## How the Files API works
2347
from line 52
2852* **Reference files** in [Messages](https://platform.claude.com/docs/en/api/messages/create) requests using the `file_id` instead of re-uploading content
2953* **Manage your files** with list, retrieve, and delete operations
3054
31<Warning id="workspace-scoped-access">
32 **Uploaded files are accessible to your entire workspace, not scoped to an end user, conversation, or session.** Any API key with access to a workspace can access any files uploaded to that workspace. Every service account, and every user whose organization role allows API access, can use the Default Workspace in addition to any workspace you add them to, so keep files that must stay separate in their own workspace and access them only with [keys scoped to that workspace](https://platform.claude.com/docs/en/manage-claude/workspaces#api-keys-and-resource-scoping). Never accept `file_id` values from end users or other untrusted sources: a user-supplied file ID would let one user of your application read content that another user uploaded. Treat file IDs as server-side references, and keep the mapping between your users and their files in your application.
55<Warning id="workspace-scoped-access" title="Uploaded files are accessible to your entire workspace">
56 Files are not scoped to an end user, conversation, or session. Any API key with access to a workspace can access any files uploaded to that workspace. Every service account, and every user whose organization role allows API access, can use the Default Workspace in addition to any workspace you add them to, so keep files that must stay separate in their own workspace and access them only with [keys scoped to that workspace](https://platform.claude.com/docs/en/manage-claude/workspaces#api-keys-and-resource-scoping). Never accept `file_id` values from end users or other untrusted sources: a user-supplied file ID would let one user of your application read content that another user uploaded. Treat file IDs as server-side references, and keep the mapping between your users and their files in your application.
3357
3458 If you are building a multitenant application on the Files API, create a separate [workspace](https://platform.claude.com/docs/en/manage-claude/workspaces) for each tenant. The workspace is the isolation boundary for files, so a workspace for each tenant gives that tenant’s data hard isolation from every other tenant. Each organization can have up to 100 workspaces; contact your account team if you need more.
3559</Warning>
from line 967
943967* Files persist until you delete them with the `DELETE /v1/files/{file_id}` endpoint or they reach their `expires_at`.
944968* Deleted files cannot be recovered.
945969* Files are inaccessible through the API shortly after deletion, but they may persist in active Messages API calls and associated tool uses.
946* Files that users delete will be deleted in accordance with Anthropic’s [data retention policy](https://privacy.claude.com/en/articles/7996866-how-long-do-you-store-my-organization-s-data). For ZDR eligibility across all features, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
970* Files that users delete will be deleted in accordance with Anthropic’s [data retention policy](https://privacy.claude.com/en/articles/7996866-how-long-do-you-store-my-organization-s-data).
947971
948972### File expiration
949973
from line 1067
10431067 Process and analyze visual input and generate text and code from images.
10441068 </Card>
10451069</CardGroup>
1070
1071## Availability
1072
1073The Files API is available on every model.
1074
1075### API support
1076
1077#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
1078
1079* Availability: Available
1080* Usage: Refer to an uploaded file by its `file_id` in a `document`, `image` or `container_upload` block.
1081
1082| Status | Cloud platforms |
1083| --------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
1084| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed), [Microsoft Foundry (hosted on Anthropic)](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
1085
1086#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
1087
1088* Availability: Available
1089* Usage: Mount uploaded files in the session’s sandbox as session resources (read-only, under `/mnt/session/uploads`), or refer to a `file_id` in an `image` or `document` block of a `user.message` event. Files the agent writes to `/mnt/session/outputs/` can be downloaded with the Files API.
1090
1091| Status | Cloud platforms |
1092| --------- | ----------------------- |
1093| Available | All supported platforms |
1094
1095### Data handling
1096
1097[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
1098
1099| Item | Zero data retention | HIPAA |
1100| ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------ |
1101| Status | Not eligible | Not eligible |
1102| Details | Files retained until explicitly deleted or they reach their configured expiration. [Learn more](https://platform.claude.com/docs/en/build-with-claude/files#data-retention) | — |
1103
1104Applicable endpoints:
1105
1106* `/v1/files`
10461107
build-with-claude/mid-conversation-system-messages Changed · +62 / -3 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling
from line 2
22title: Mid-conversation system messages and tool changes
33url: https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages
44description: Change system instructions or tool availability partway through a conversation without invalidating the cached prefix that came before them.
5featureMetadata:
6 status: ga
7 zdr:
8 eligibility: eligible
9 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). Request-shape capability of the Messages API; mid-conversation system messages flow through the standard inference path and nothing is stored server-side after the response.
10 supportedModels:
11 - claude-fable-5-1
12 - claude-mythos-5-1
13 - claude-fable-5
14 - claude-mythos-5
15 - claude-mythos-preview
16 - claude-opus-5-5
17 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-haiku-5-5
20 - claude-opus-4-8
21 supportedPlatforms:
22 Claude API: ga
23 Claude Platform on AWS: ga
24 Amazon Bedrock: ga
25 Claude Models on Google Cloud (Google managed): ga
526---
627
7<Note>
8 To learn how zero data retention (ZDR) applies to this feature, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
9</Note>
28**API support**: Usage differs by product.
1029
30* **Messages API**: Append a message with `role` `"system"` to `messages` at the point where the instruction starts to apply.
31* **Managed Agents**: Send a `system.message` event. It applies to that turn and all later turns; the agent’s own `system` field cannot change during a session.
32
1133System instructions normally live in the top-level `system` field, ahead of every message in the conversation. That position is great for [prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching): the system prompt is part of the stable prefix, so subsequent turns hit the cache. It is a poor position for instructions you only discover you need partway through a session, because editing the top-level `system` field changes the very beginning of the prompt and invalidates the cache for everything that follows.
1234
1335Mid-conversation system messages close that gap. You append a `{"role": "system"}` message at the point in the conversation where the new instruction becomes relevant, instead of editing the top-level `system` field. The cached prefix stays the same, so the next request still reads it from cache, and the new instruction is still applied as a system instruction rather than as ordinary user text.
from line 2385
23632385 How `tool_use` and `tool_result` blocks are structured in the `messages` array.
23642386 </Card>
23652387</CardGroup>
2388
2389## Availability
2390
2391Mid-conversation system messages are available on select models and eligible for zero data retention and HIPAA readiness.
2392
2393### API support
2394
2395#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
2396
2397* Availability: Available
2398* Usage: Append a message with `role` `"system"` to `messages` at the point where the instruction starts to apply.
2399
2400| Status | Cloud platforms | Models |
2401| --------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
2402| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview) |
2403
2404#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
2405
2406* Availability: Available
2407* Usage: Send a `system.message` event. It applies to that turn and all later turns; the agent’s own `system` field cannot change during a session.
2408
2409| Status | Cloud platforms | Models |
2410| --------- | ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
2411| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview) |
2412
2413### Data handling
2414
2415[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
2416
2417| Item | Zero data retention | HIPAA |
2418| ------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
2419| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
2420| Details | Request-shape capability of the Messages API; mid-conversation system messages flow through the standard inference path and nothing is stored server-side after the response. | — |
2421
2422Applicable endpoints:
2423
2424* `/v1/messages` with `role: "system"` messages
23662425
build-with-claude/multilingual-support Changed · +49 / -0 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
from line 2
22title: Multilingual support
33url: https://platform.claude.com/docs/en/build-with-claude/multilingual-support
44description: Claude excels at tasks across multiple languages, maintaining strong cross-lingual performance relative to English.
5featureMetadata:
6 status: ga
7 supportedModels:
8 - claude-fable-5-1
9 - claude-mythos-5-1
10 - claude-fable-5
11 - claude-mythos-5
12 - claude-mythos-preview
13 - claude-opus-5-5
14 - claude-opus-5
15 - claude-sonnet-5-5
16 - claude-sonnet-5
17 - claude-haiku-5-5
18 - claude-haiku-4-5-20251001
19 - claude-opus-4-8
20 - claude-opus-4-7
21 - claude-opus-4-6
22 - claude-sonnet-4-6
23 - claude-opus-4-5-20251101
24 - claude-sonnet-4-5-20250929
25 supportedPlatforms:
26 Claude API: ga
27 Claude Platform on AWS: ga
28 Amazon Bedrock: ga
29 Amazon Bedrock (InvokeModel): ga
30 Claude Models on Google Cloud (Google managed): ga
31 Microsoft Foundry: ga
532---
633
734## Overview
from line 241
214241 Evaluate translation and localization quality before you ship.
215242 </Card>
216243</CardGroup>
244
245## Availability
246
247Multilingual support is available on every model.
248
249### API support
250
251#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
252
253* Availability: Always on
254
255| Status | Cloud platforms |
256| --------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
257| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
258
259#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
260
261* Availability: Always on
262
263| Status | Cloud platforms |
264| --------- | ------------------------------------------------------------------------------------------------------------------ |
265| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) |
217266
build-with-claude/overview Changed · +105 / -97 lines
## Platform features ## Messages API features ## Managed Agents features ## Feature availability ## Model capabilities ## Tools ### Server-side tools ### Client-side tools ## Tool infrastructure ## Context management ## Files and assets
from line 4
44description: Explore Claude's advanced features and capabilities.
55---
66
7Claude’s API surface is organized into five areas:
7For each feature’s eligibility under zero data retention (ZDR) and HIPAA readiness, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility).
88
9* **Model capabilities:** Control how Claude reasons and formats responses.
10* **Tools:** Let Claude take actions on the web or in your environment.
11* **Tool infrastructure:** Handle discovery and orchestration at scale.
12* **Context management:** Keep long-running sessions efficient.
13* **Files and assets:** Manage the documents and data you provide to Claude.
14
15If you’re new, start with [model capabilities](https://platform.claude.com/docs/en/build-with-claude/overview#model-capabilities) and [tools](https://platform.claude.com/docs/en/build-with-claude/overview#tools). Return to the other sections when you’re ready to optimize cost, latency, or scale.
16
179For administration and governance, see the [Admin API](https://platform.claude.com/docs/en/manage-claude/admin-api), the [Usage and Cost API](https://platform.claude.com/docs/en/manage-claude/usage-cost-api), and the [Compliance API](https://platform.claude.com/docs/en/manage-claude/compliance-api).
1810
19## Feature availability
11## Platform features
2012
21The Availability column in each of the following tables lists the platforms that offer a feature. A platform listed without a label offers the feature as stable, fully supported, and recommended for production use, with no beta header and with standard API [versioning](https://platform.claude.com/docs/en/api/versioning) guarantees. A label after a platform name marks one of the following classifications on that platform. Not all features pass through every stage, and a feature may enter at any stage or skip stages.
13Features of the Claude Platform that work across its APIs. Each feature’s page says where you can use it.
2214
23| Classification | Description |
24| -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
25| **Beta** | Preview features used for gathering feedback and iterating on a less mature use case. Availability may be limited, including through sign-up requirements or waitlists, and may not be publicly announced.<br /><br />Features may change significantly or be discontinued based on feedback. Not guaranteed for ongoing production use. Breaking changes are possible with notice, and some platform-specific limitations may apply. Most beta features on the Claude API and [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) have a [beta header](https://platform.claude.com/docs/en/api/beta-headers). |
26| **Deprecated** | Feature is still functional but no longer recommended. A migration path and removal timeline are provided. |
27| **Retired** | Feature is no longer available. |
28
29**Platform labels:** Claude API (Anthropic first-party) · [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock) (AWS-operated) · [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) (Anthropic-operated on AWS) · [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed) (Google-operated) · [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) (Anthropic-operated on Azure)
30
31## Model capabilities
32
33Ways to steer Claude and Claude’s direct outputs, including response format, reasoning depth, and input modalities.
34
3515<Tip>
3616 You can discover which capabilities a model supports programmatically. The [Models API](https://platform.claude.com/docs/en/api/models/list) returns `max_input_tokens`, `max_tokens`, and a `capabilities` object for every available model.
3717</Tip>
3818
39The ZDR column indicates whether a feature is available under a Zero Data Retention arrangement. For most features this depends only on what the feature mechanism retains; for features tied to specific models, model-level ZDR availability also applies. See [Model-specific data retention requirements](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
19| Feature | Description |
20| :------------------------------------------------------------------------------------------------------------------------------------------------------ | :------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
21| [Advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool) | Lets a faster executor model consult a more capable advisor model mid-generation for a plan or a course correction. |
22| [Agent Skills](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview) | Reusable packages of instructions, metadata and optional scripts that Claude uses automatically when they are relevant to a task. |
23| ↳ [Custom Skills](https://platform.claude.com/docs/en/build-with-claude/skills-guide#managing-custom-skills) | Skills you write and upload to your workspace. |
24| ↳ [Pre-built Agent Skills](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview#pre-built-agent-skills) | Anthropic’s pre-built skills for PowerPoint, Excel, Word and PDF files. |
25| [Batch processing](https://platform.claude.com/docs/en/build-with-claude/batch-processing) | Processes many Messages API requests asynchronously in one batch, at half the standard price. |
26| [Browser and computer use SDK toolsets](https://platform.claude.com/docs/en/agents-and-tools/tool-use/browser-use-sdk) | Classes in the Python and TypeScript SDKs that connect the browser use or computer use tool to a browser or desktop that you supply. |
27| [Browser use tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/browser-use-tool) | Lets Claude navigate, read, and interact with webpages in your own browser environment. |
28| [Citations](https://platform.claude.com/docs/en/build-with-claude/citations) | Claude’s answers about your documents cite the exact passages that support each claim. |
29| [Classifier refusals](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback) | Safety classifiers on some models can decline a request; the decline comes back as a normal response, not an error. |
30| ↳ [Client-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#client-side-fallback) | A middleware in the Claude SDKs that retries a refused request on the fallback models you list. |
31| ↳ [Fallback credit](https://platform.claude.com/docs/en/build-with-claude/fallback-credit) | Bills a retry of a refused request on another model as if the conversation had been on that model from the start, so you do not pay the prompt-cache write twice. |
32| ↳ [Server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#server-side-fallback) | The API retries a refused request on a fallback model within the same call and returns one response. |
33| [Computer use tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool) | Gives Claude screenshot, mouse, and keyboard control of a desktop environment. |
34| [Custom tools](https://platform.claude.com/docs/en/agents-and-tools/tool-use/define-tools) | Tools you define with a name, description and input schema. Claude asks to call them; your code runs them and returns the result. |
35| [Data residency](https://platform.claude.com/docs/en/manage-claude/data-residency) | Controls for where model inference runs (inference geo) and where data is stored at rest (workspace geo). |
36| [Effort](https://platform.claude.com/docs/en/build-with-claude/effort) | A setting that controls how many tokens Claude spends on a response, trading thoroughness for token efficiency. |
37| ↳ [Per-message effort](https://platform.claude.com/docs/en/build-with-claude/effort#change-effort-mid-conversation-beta) | Changing the effort level for later turns of a conversation with a per-message `output_config`, which keeps the prompt cache. |
38| [Fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) | Up to 2.5 times more output tokens per second from supported models, at premium pricing. |
39| [Files API](https://platform.claude.com/docs/en/build-with-claude/files) | Upload a file once and refer to it by `file_id` in later requests, and download files that skills or the code execution tool create. |
40| [MCP connector](https://platform.claude.com/docs/en/agents-and-tools/mcp-connector) | Connect Claude to remote MCP servers so it can call their tools, without running your own MCP client. |
41| ↳ [Client-side MCP helpers](https://platform.claude.com/docs/en/agents-and-tools/mcp-connector#client-side-mcp-helpers) | SDK helpers that convert between MCP types and Claude API types, for when you run your own MCP client, for example for local servers, prompts or resources. |
42| [MCP tunnels](https://platform.claude.com/docs/en/agents-and-tools/mcp-tunnels/overview) | Connect Claude to MCP servers inside your private network over an outbound-only connection, with no inbound ports open. |
43| [Mid-conversation system messages](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages) | A system instruction added partway through a conversation, without changing the top-level system prompt or the cached prefix before it. |
44| ↳ [Mid-conversation tool changes](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages#mid-conversation-tool-changes) | Offer or withdraw tools partway through a conversation without changing the tools array, so the cached prefix stays valid. |
45| ↳ [Turn-scoped system messages](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages#turn-scoped-system-messages) | A mid-conversation system message that applies only until the next user message, then stays in the history but renders nothing and costs no input tokens. |
46| [Multilingual support](https://platform.claude.com/docs/en/build-with-claude/multilingual-support) | Claude works in many languages, with strong performance relative to English. |
47| [OpenTelemetry](https://platform.claude.com/docs/en/manage-claude/opentelemetry) | Traces your Messages API requests, tool uses, and Managed Agents sessions with OpenTelemetry spans from the Claude SDK and the Claude API. |
48| ↳ [API tracing](https://platform.claude.com/docs/en/manage-claude/opentelemetry/api) | OpenTelemetry spans that the Claude API sends to your observability platform for each model request and tool use that it runs. |
49| ↳ [SDK tracing](https://platform.claude.com/docs/en/manage-claude/opentelemetry/sdk) | OpenTelemetry spans that the Claude SDK emits in your application for each request it sends and each tool that its tool runner runs. |
50| [Parallel tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/parallel-tool-use) | Claude can call several tools in one response. |
51| [PDF support](https://platform.claude.com/docs/en/build-with-claude/pdf-support) | Send PDFs to Claude so it can read the text, pictures, charts and tables in them. |
52| [Programmatic tool calling](https://platform.claude.com/docs/en/agents-and-tools/tool-use/programmatic-tool-calling) | Lets Claude call your tools from code in the code execution container, instead of one model round trip for each call. |
53| [Prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) | Reuses a cached prompt prefix from a recent request to reduce processing time and cost. |
54| ↳ [1-hour duration](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#1-hour-cache-duration) | A 1-hour prompt cache lifetime, for a higher cache write price than the 5-minute default. |
55| ↳ [5-minute duration (default)](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#how-prompt-caching-works) | The default prompt cache lifetime. Each use of the cached content refreshes it at no extra cost. |
56| ↳ [Automatic caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#automatic-caching) | Prompt caching with one top-level `cache_control` field: the cache breakpoint goes on the last cacheable block and moves forward as the conversation grows. |
57| ↳ [Cache diagnostics](https://platform.claude.com/docs/en/build-with-claude/cache-diagnostics) | Compares a request with your previous one and reports where the prompt prefix changed, to explain cache misses. |
58| ↳ [Explicit cache breakpoints](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#explicit-cache-breakpoints) | Prompt caching with `cache_control` on individual content blocks, to choose exactly what is cached. |
59| [SDK middleware](https://platform.claude.com/docs/en/cli-sdks-libraries/middleware) | A hook in the Claude SDKs that runs your code before each request is sent and after its response is received. |
60| [Search results](https://platform.claude.com/docs/en/build-with-claude/search-results) | Content blocks that hold your own search results, with a source and title, so Claude can cite them the same way it cites web search results. |
61| [Service tiers](https://platform.claude.com/docs/en/api/service-tiers) | Lets a request use your organization’s Priority Tier capacity. |
62| [Session budgets](https://platform.claude.com/docs/en/managed-agents/budgets) | A hard spending limit in US dollars on one session, measured at public list prices. |
63| [Streaming refusals](https://platform.claude.com/docs/en/test-and-evaluate/strengthen-guardrails/handle-streaming-refusals) | A streamed response ends with `stop_reason` `"refusal"` when streaming classifiers stop it for a possible policy violation. |
64| [Structured outputs](https://platform.claude.com/docs/en/build-with-claude/structured-outputs) | Constrains Claude’s response to a JSON schema that you give, so the output is valid and parseable. |
65| [Task budgets](https://platform.claude.com/docs/en/build-with-claude/task-budgets) | Tells Claude how many tokens it has for a full agentic loop, so it can plan its work and finish before the budget runs out. |
66| [Thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) | Claude works through a problem in thinking blocks before it writes its answer. |
67| ↳ [Adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking-steering-and-cost) | Thinking mode in which Claude decides, for each request, whether to think and how much. |
68| ↳ [Extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) | Manual thinking mode in which you set a fixed thinking token budget (`budget_tokens`) on each request. |
69| ↳ [Preserved thinking](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking) | Claude uses thinking blocks from earlier turns only if the current model can read them and nothing before them (system prompt, tools, earlier messages) has changed. |
70| [Threshold compaction](https://platform.claude.com/docs/en/build-with-claude/compaction-threshold) | Automatic compaction: when a conversation reaches a token threshold, the API summarizes older context inside the request. |
71| [Token counting](https://platform.claude.com/docs/en/build-with-claude/token-counting) | Counts the tokens in a message before you send it to Claude. |
72| [Tool runner](https://platform.claude.com/docs/en/agents-and-tools/tool-use/tool-runner) | An SDK helper that runs the tool-use loop for you: it runs your tools when Claude calls them and sends the results back. |
73| [Tool search tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/tool-search-tool) | Lets Claude search your tool catalog and load only the tool definitions it needs. |
74| [Vision (images)](https://platform.claude.com/docs/en/build-with-claude/vision) | Claude can read and analyze images that you send to it. |
75| [Web domain restrictions](https://platform.claude.com/docs/en/managed-agents/tools-web-restrictions) | Lists of allowed or blocked domains that control which sites the web search and web fetch tools can reach. |
76| [Web fetch tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool) | A server tool that lets Claude get the full content of web pages and PDF documents at given URLs. |
77| [Web search tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool) | A server tool that lets Claude search the web for current information. |
4078
41| Feature | Description | Zero Data Retention (ZDR) | Availability |
42| --------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------- |
43| [Context windows](https://platform.claude.com/docs/en/build-with-claude/context-windows) | Up to 1M tokens for processing large documents, extensive code bases, and long conversations. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
44| [Adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) | Let Claude dynamically determine when and how much to think. The only thinking mode on Claude 4.7 and later models. Use the `effort` parameter to control thinking depth. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
45| [Batch processing](https://platform.claude.com/docs/en/build-with-claude/batch-processing) | Process large volumes of requests asynchronously for cost savings. Send batches with a large number of queries per batch. Batch API calls cost 50% less than standard API calls. | Not ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws /> |
46| [Citations](https://platform.claude.com/docs/en/build-with-claude/citations) | Ground Claude’s responses in source documents. With Citations, Claude can provide detailed references to the exact sentences and passages it uses to generate responses, leading to more verifiable, trustworthy outputs. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
47| [Data residency](https://platform.claude.com/docs/en/manage-claude/data-residency) | Control where model inference runs using geographic controls. Specify `"global"` or `"us"` routing per request through the `inference_geo` parameter. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws /> |
48| [Effort](https://platform.claude.com/docs/en/build-with-claude/effort) | Control how many tokens Claude uses when responding with the `effort` parameter, trading off between response thoroughness and token efficiency. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
49| [Fallback credit](https://platform.claude.com/docs/en/build-with-claude/fallback-credit) | Avoid paying the prompt-cache cost twice when you retry a refused request on another model. The refusal carries a credit token, and echoing it on the retry bills the retry as though the conversation had been on the new model from the start. Message Batches results do not include fallback credit tokens. | Not ZDR eligible\* | <PlatformAvailability claudeApiBeta claudePlatformAwsBeta bedrockBeta vertexAiBeta azureAiBeta /> |
50| [PDF support](https://platform.claude.com/docs/en/build-with-claude/pdf-support) | Process and analyze text and visual content from PDF documents. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
51| [Search results](https://platform.claude.com/docs/en/build-with-claude/search-results) | Get natural citations for RAG applications by providing search results with proper source attribution. Achieve web search-quality citations for custom knowledge bases and tools. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
52| [Server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback) | Retry a refused request inside a single API call. Use the `"default"` mode to apply Anthropic’s recommended fallback models, or name up to three models of your own; when the requested model declines, the API runs the next model in the chain on the same request. The `fallbacks` parameter is not available in the Message Batches API. | Not ZDR eligible\* | <PlatformAvailability claudeApiBeta /> |
53| [Structured outputs](https://platform.claude.com/docs/en/build-with-claude/structured-outputs) | Guarantee schema conformance with two approaches: JSON outputs for structured data responses, and strict tool use for validated tool inputs. | [ZDR eligible (qualified)](https://platform.claude.com/docs/en/build-with-claude/structured-outputs#data-retention)\* | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
54| [Thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) | Enhanced reasoning capabilities for complex tasks, with an optional summary of Claude’s reasoning before its final answer. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
79## Messages API features
5580
56## Tools
81| Feature | Description |
82| :------------------------------------------------------------------------------------------------------------------------------ | :------------------------------------------------------------------------------------------------------------------------------------------- |
83| [Apple Foundation Models](https://platform.claude.com/docs/en/cli-sdks-libraries/libraries/apple-foundation-models) | A Swift package that makes Claude available as a language model in Apple’s Foundation Models framework. |
84| [Bash tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/bash-tool) | Lets Claude request shell commands, which run in a bash session that keeps its state between commands. |
85| [Code execution tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/code-execution-tool) | Lets Claude run Python and Bash code in a sandboxed container. |
86| [Context editing](https://platform.claude.com/docs/en/build-with-claude/context-editing) | Clears selected content, such as old tool results or old thinking blocks, from the conversation history as it grows. |
87| [Forced tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/define-tools#forcing-tool-use) | Makes Claude call a tool, or one specific tool, instead of answering directly. |
88| [Memory tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/memory-tool) | Lets Claude store and read information across conversations in a directory of memory files. |
89| ↳ [Memory tool helpers](https://platform.claude.com/docs/en/agents-and-tools/tool-use/memory-tool#implement-the-memory-handler) | SDK helpers that handle the memory tool’s interface and tool-use loop, so your code supplies only the storage. |
90| [Message streaming](https://platform.claude.com/docs/en/build-with-claude/streaming) | Receiving a Messages API response incrementally, as server-sent events, while it is generated. |
91| ↳ [Fine-grained tool streaming](https://platform.claude.com/docs/en/agents-and-tools/tool-use/fine-grained-tool-streaming) | Streams a tool’s input to your client as Claude writes it, without the server first buffering and checking the JSON. |
92| [On-demand compaction](https://platform.claude.com/docs/en/build-with-claude/compaction-on-demand) | Replaces the earlier messages of a conversation with a summary that Claude writes when your application asks for it. |
93| ↳ [Background compaction](https://platform.claude.com/docs/en/build-with-claude/compaction-background) | Runs the on-demand compaction request while the conversation continues on its full history, then swaps in the summary when it arrives. |
94| ↳ [Compaction and preserved thinking](https://platform.claude.com/docs/en/build-with-claude/compaction-thinking-blocks) | Thinking blocks in the turns you keep after on-demand compaction can stay valid, under stated conditions, on models with preserved thinking. |
95| ↳ [Keep-tail compaction](https://platform.claude.com/docs/en/build-with-claude/compaction-keep-recent-turns) | On-demand compaction that summarizes the older turns and keeps the last few turns word for word after the summary. |
96| ↳ [Tool runner compaction helper](https://platform.claude.com/docs/en/build-with-claude/compaction-on-demand#compact-in-a-loop) | A method on the SDK tool runner that sends the compaction request for you once the current turn and its tool calls finish. |
97| [OpenAI SDK compatibility](https://platform.claude.com/docs/en/cli-sdks-libraries/libraries/openai-sdk) | A compatibility layer that lets you call the Claude API with the OpenAI SDK, meant for testing and comparing models. |
98| [Strict tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/strict-tool-use) | Makes Claude’s tool inputs always match the tool’s JSON Schema. |
99| [Text editor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/text-editor-tool) | Lets Claude view, create, and edit text files with an Anthropic-defined tool. |
100| [Typed structured output helpers](https://platform.claude.com/docs/en/build-with-claude/structured-outputs#usage) | SDK helpers that build the JSON schema from a type in your code and parse Claude’s response into that type. |
57101
58Built-in tools that Claude calls through `tool_use`. Server-side tools are run by the platform; client-side tools are implemented and run by you.
102## Managed Agents features
59103
60### Server-side tools
61
62| Feature | Description | ZDR | Availability |
63| --------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------- | --------------------------------------------------------------------- |
64| [Advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool) | Pair a faster executor model with a higher-intelligence advisor model that provides strategic guidance mid-generation for long-horizon agentic workloads. | ZDR eligible | <PlatformAvailability claudeApiBeta claudePlatformAwsBeta /> |
65| [Code execution](https://platform.claude.com/docs/en/agents-and-tools/tool-use/code-execution-tool) | Run code in a sandboxed environment for advanced data analysis, calculations, and file processing. Free when used with web search or web fetch. | Not ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws azureAi />† |
66| [Web fetch](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool) | Retrieve full content from specified webpages and PDF documents for in-depth analysis. | ZDR eligible\* | <PlatformAvailability claudeApi claudePlatformAws azureAi /> |
67| [Web search](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool) | Augment Claude’s comprehensive knowledge with current, real-world data from across the web. | ZDR eligible\* | <PlatformAvailability claudeApi claudePlatformAws vertexAi azureAi /> |
68
69### Client-side tools
70
71| Feature | Description | ZDR | Availability |
72| ----------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------ | ----------------------------------------------------------------------------------------- |
73| [Bash](https://platform.claude.com/docs/en/agents-and-tools/tool-use/bash-tool) | Run bash commands and scripts to interact with the system shell and perform command-line operations. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
74| [Browser use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/browser-use-tool) | Navigate, read, and interact with webpages in your own browser environment. | ZDR eligible | <PlatformAvailability claudeApi vertexAi /> |
75| [Computer use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool) | Control computer interfaces by taking screenshots and issuing mouse and keyboard commands. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAwsBeta bedrockBeta vertexAi azureAiBeta /> |
76| [Memory](https://platform.claude.com/docs/en/agents-and-tools/tool-use/memory-tool) | Allow Claude to store and retrieve information across conversations. Build knowledge bases over time, maintain project context, and learn from past interactions. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
77| [Text editor](https://platform.claude.com/docs/en/agents-and-tools/tool-use/text-editor-tool) | Create and edit text files with a built-in text editor interface for file manipulation tasks. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
78
79## Tool infrastructure
80
81Infrastructure that supports discovering, orchestrating, and scaling tool use.
82
83| Feature | Description | ZDR | Availability |
84| ------------------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------- | ----------------------------------------------------------------------------- |
85| [Agent Skills](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/overview) | Extend Claude’s capabilities with Skills. Use pre-built Skills (PowerPoint, Excel, Word, PDF) or create custom Skills with instructions and scripts. Skills use progressive disclosure to efficiently manage context. | Not ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws azureAi />†‡ |
86| [Fine-grained tool streaming](https://platform.claude.com/docs/en/agents-and-tools/tool-use/fine-grained-tool-streaming) | Stream tool use parameters without buffering/JSON validation, reducing latency for receiving large parameters. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
87| [MCP connector](https://platform.claude.com/docs/en/agents-and-tools/mcp-connector) | Connect to remote [MCP](https://modelcontextprotocol.io/) servers directly from the Messages API without a separate MCP client. | Not ZDR eligible | <PlatformAvailability claudeApiBeta claudePlatformAwsBeta azureAiBeta /> |
88| [Programmatic tool calling](https://platform.claude.com/docs/en/agents-and-tools/tool-use/programmatic-tool-calling) | Allow Claude to call your tools programmatically from within code execution containers, reducing latency and token consumption for multi-tool workflows. | Not ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws azureAi />† |
89| [Tool search](https://platform.claude.com/docs/en/agents-and-tools/tool-use/tool-search-tool) | Scale to thousands of tools by dynamically discovering and loading tools on demand using regex- and BM25-based search, optimizing context usage and improving tool selection accuracy. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
90
91## Context management
92
93Infrastructure for controlling and optimizing Claude’s context window.
94
95| Feature | Description | ZDR | Availability |
96| ------------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------ | ------------------------------------------------------------------------------------------------- |
97| [Compaction at a token threshold](https://platform.claude.com/docs/en/build-with-claude/compaction-threshold) | Server-side context summarization for long-running conversations. When input tokens reach the trigger threshold, the API automatically summarizes earlier parts of the conversation. | ZDR eligible | <PlatformAvailability claudeApiBeta claudePlatformAwsBeta bedrockBeta vertexAiBeta azureAiBeta /> |
98| [Context editing](https://platform.claude.com/docs/en/build-with-claude/context-editing) | Automatically manage conversation context with configurable strategies. Supports clearing tool results when approaching token limits and managing thinking blocks in extended thinking conversations. | ZDR eligible | <PlatformAvailability claudeApiBeta claudePlatformAwsBeta bedrockBeta vertexAiBeta azureAiBeta /> |
99| [Automatic prompt caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#automatic-caching) | Simplify prompt caching to a single API parameter. The system automatically caches the last cacheable block in your request, moving the cache point forward as conversations grow. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
100| [Prompt caching (5m)](https://platform.claude.com/docs/en/build-with-claude/prompt-caching) | Provide Claude with more background knowledge and example outputs to reduce costs and latency. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
101| [Prompt caching (1h)](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#1-hour-cache-duration) | Extended 1-hour cache duration for less frequently accessed but important context, complementing the standard 5-minute cache. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
102| [Token counting](https://platform.claude.com/docs/en/build-with-claude/token-counting) | Token counting allows you to determine the number of tokens in a message before sending it to Claude, helping you make informed decisions about your prompts and usage. | ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws bedrock vertexAi azureAi /> |
103
104## Files and assets
105
106Manage files and assets for use with Claude.
107
108| Feature | Description | ZDR | Availability |
109| ------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------- | ---------------- | ------------------------------------------------------------- |
110| [Files API](https://platform.claude.com/docs/en/build-with-claude/files) | Upload and manage files to use with Claude without re-uploading content with each request. Supports PDFs, images, and text files. | Not ZDR eligible | <PlatformAvailability claudeApi claudePlatformAws azureAi />† |
111
112\* **Structured outputs:** Your prompts and Claude’s outputs are not stored. Only JSON schemas are cached, for up to 24 hours since last use. **Web search and web fetch:** ZDR-eligible except when [dynamic filtering](https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-search-tool#dynamic-filtering) is enabled. **Fallback credit and server-side fallback:** The features retain no message content, but they handle refusals from the Claude Fable models, which [are not available under ZDR](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). See [ZDR details](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility).
113
114† On Microsoft Foundry, feature availability differs by [hosting option](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry#hosting-options). These features are available on Hosted on Anthropic deployments, and not on Hosted on Azure deployments.
115
116‡ On Microsoft Foundry, the [Skill version download endpoint](https://platform.claude.com/docs/en/api/beta/skills/versions/download) (`GET /v1/skills/{skill_id}/versions/{version}/content`) is not supported.
104| Feature | Description |
105| :-------------------------------------------------------------------------------------------------------- | :----------------------------------------------------------------------------------------------------------------------------------------------------------- |
106| [Agent configurations](https://platform.claude.com/docs/en/managed-agents/agent-setup) | A reusable, versioned configuration of the model, system prompt, tools, MCP servers and skills that a session runs. |
107| [Agent toolset](https://platform.claude.com/docs/en/managed-agents/tools) | A set of built-in tools (bash, read, write, edit, glob, grep, web fetch and web search) that the agent uses in its sandbox. |
108| [Cloud sandboxes](https://platform.claude.com/docs/en/managed-agents/environments) | Isolated Linux containers on Anthropic-managed infrastructure where a session’s tools and code run. |
109| [Console agent builder](https://platform.claude.com/docs/en/managed-agents/onboarding) | A visual interface in the Claude Console to create, configure and test agents. |
110| [Dreams](https://platform.claude.com/docs/en/managed-agents/dreams) | A job that reads a memory store and past session transcripts and writes a new, reorganized memory store. |
111| [Event deltas](https://platform.claude.com/docs/en/managed-agents/event-deltas) | Live previews of the agent’s response text on the session event stream while the model is still generating it. |
112| [GitHub access](https://platform.claude.com/docs/en/managed-agents/github) | Access to GitHub from a session: repositories mounted in the sandbox, and the GitHub MCP server for pull requests. |
113| [Memory stores](https://platform.claude.com/docs/en/managed-agents/memory) | Collections of text documents that an agent reads and writes, so that information persists across sessions. |
114| [Multiagent orchestration](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration) | Lets the agent of a session hand work to other agents, as subagents or in workflows, and consult an advisor model. |
115| ↳ [Dynamic workflows](https://platform.claude.com/docs/en/managed-agents/workflow-runs) | Programs that an agent writes to run many agents in the background and combine their results. |
116| ↳ [Session threads](https://platform.claude.com/docs/en/managed-agents/session-threads) | The thread in which each agent of a multiagent session works, which you can list, stream, interrupt and archive. |
117| [Outcomes](https://platform.claude.com/docs/en/managed-agents/define-outcomes) | A target result for a session, with a rubric that a separate grader uses to check the agent’s work until the target is met. |
118| [Permission policies](https://platform.claude.com/docs/en/managed-agents/permission-policies) | Settings that control whether the agent’s built-in tools and MCP tools run automatically, wait for your approval, or are checked by the server on each call. |
119| [Scheduled deployments](https://platform.claude.com/docs/en/managed-agents/scheduled-deployments) | Runs an agent on a cron schedule, so that the agent starts sessions by itself. |
120| [Self-hosted sandboxes](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes) | Sandboxes on your own infrastructure that run the agent’s tool calls, while Claude still runs on Anthropic’s side. |
121| ↳ [Self-hosted workers](https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes-workers) | A process that you run on your own infrastructure to claim sessions from a self-hosted environment’s queue and run their tool calls. |
122| [Session viewer](https://platform.claude.com/docs/en/managed-agents/session-observability) | A session viewer in the Claude Console and per-session usage totals, to see what an agent did and what it used. |
123| [Vaults](https://platform.claude.com/docs/en/managed-agents/vaults) | Stores of per-user credentials for third-party services that you register once and reference by ID when you create a session. |
124| [Webhooks](https://platform.claude.com/docs/en/managed-agents/webhooks) | Notifications to your HTTPS endpoint when events happen in your Managed Agents resources, such as a change in a session’s status. |
117125
build-with-claude/pdf-support Changed · +78 / -8 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling
from line 7
77 zdr:
88 eligibility: eligible
99 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
10 supportedModels:
11 - claude-fable-5-1
12 - claude-mythos-5-1
13 - claude-fable-5
14 - claude-mythos-5
15 - claude-mythos-preview
16 - claude-opus-5-5
17 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
21 - claude-haiku-4-5-20251001
22 - claude-opus-4-8
23 - claude-opus-4-7
24 - claude-opus-4-6
25 - claude-sonnet-4-6
26 - claude-opus-4-5-20251101
27 - claude-sonnet-4-5-20250929
1028 supportedPlatforms:
1129 Claude API: ga
1230 Claude Platform on AWS: ga
13 Amazon Bedrock: ga
14 Claude Models on Google Cloud (Google managed): ga
15 Microsoft Foundry: ga
31 Claude Models on Google Cloud (Anthropic managed): ga
32 Amazon Bedrock:
33 availability: ga
34 note: Base64 sources only.
35 Amazon Bedrock (InvokeModel):
36 availability: ga
37 note: Base64 sources only. With the Converse API, visual PDF analysis needs citations turned on.
38 Claude Models on Google Cloud (Google managed):
39 availability: ga
40 note: Base64 sources only.
41 Microsoft Foundry:
42 availability: ga
43 note: "`file_id` sources need a Hosted on Anthropic deployment."
1644---
1745
1846You can ask Claude about any text, pictures, charts, and tables in PDFs you provide. Some sample use cases:
from line 56
2856
2957Claude works with any standard PDF. Ensure your request size meets these requirements:
3058
31| Requirement | Limit |
32| ------------------------- | --------------------------------------------------------------------------------------------------------------------------- |
33| Maximum request size | 32 MB (varies by platform; see [request size limits](https://platform.claude.com/docs/en/api/overview#request-size-limits)) |
34| Maximum pages per request | 600 (100 when the request’s context window is under 1M tokens) |
35| Format | Standard PDF (no passwords/encryption) |
59| Requirement | Limit |
60| ------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------- |
61| Maximum request size | 128 MB on the first-party Claude API (varies by platform; see [request size limits](https://platform.claude.com/docs/en/api/overview#request-size-limits)) |
62| Maximum pages per request | 600 (100 when the request’s context window is under 1M tokens) |
63| Format | Standard PDF (no passwords/encryption) |
3664
3765Both limits are on the entire request payload, including any other content sent alongside PDFs. For large PDFs, consider uploading with the [Files API](https://platform.claude.com/docs/en/build-with-claude/files) and referencing by `file_id` to keep request payloads small.
3866
from line 2016
19882016 See complete API documentation for PDF support.
19892017 </Card>
19902018</CardGroup>
2019
2020## Availability
2021
2022PDF support is available on every model and eligible for zero data retention and HIPAA readiness.
2023
2024### API support
2025
2026#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
2027
2028* Availability: Always on
2029
2030| Status | Cloud platforms |
2031| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
2032| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed) |
2033| Available<sup>1</sup> | [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
2034| Available<sup>2</sup> | [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy) |
2035| Available<sup>3</sup> | [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed) |
2036
20371. `file_id` sources need a Hosted on Anthropic deployment.
20382. Base64 sources only. With the Converse API, visual PDF analysis needs citations turned on.
20393. Base64 sources only.
2040
2041#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
2042
2043* Availability: Always on
2044
2045| Status | Cloud platforms |
2046| --------- | ----------------------- |
2047| Available | All supported platforms |
2048
2049### Data handling
2050
2051[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
2052
2053| Item | Zero data retention | HIPAA |
2054| ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- |
2055| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
2056| Details | — | HIPAA eligibility applies to PDFs sent inline through the Messages API, not through the Files API. |
2057
2058Applicable endpoints:
2059
2060* `/v1/messages`
19912061
build-with-claude/preserved-thinking Changed · +43 / -0 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
from line 2
22title: Preserved thinking
33url: https://platform.claude.com/docs/en/build-with-claude/preserved-thinking
44description: Preserved thinking lets a model use a thinking block from an earlier turn only if that model or one of a fixed set of other models produced it and nothing before the block has changed.
5featureMetadata:
6 status: ga
7 supportedModels:
8 - claude-fable-5-1
9 - claude-mythos-5-1
10 - claude-opus-5-5
11 - claude-sonnet-5-5
12 - claude-haiku-5-5
13 supportedPlatforms:
14 Claude API: ga
15 Claude Platform on AWS: ga
16 Claude Models on Google Cloud (Anthropic managed): ga
17 Claude Models on Google Cloud (Google managed): ga
18 Microsoft Foundry: ga
519---
620
21**API support**: Usage differs by product.
22
23* **Messages API**: Always on. Keep requests append-only: send earlier turns back exactly as returned, and do not change `system`, `tools` or earlier messages.
24* **Managed Agents**: Always on. Managed Agents builds the requests.
25
726Preserved thinking is a property of newer Claude models that guards against distillation. It determines whether the model can use a thinking block that you send back from an earlier turn. Starting with Claude Fable 5.1, when a `thinking` or `redacted_thinking` block comes back in a request, the API checks the block’s `signature` for two things:
827
928* **The model can read the block.** Each model reads its own thinking blocks and those of a fixed set of other models. Claude Fable 5.1 reads blocks from Claude Opus 5 and, on the Claude API, from Claude Opus 5.5; neither Claude Opus 5 nor Claude Opus 5.5 reads blocks from Claude Fable 5.1. If the current model can’t read a block, the API drops it from that request without an error. See [Switching models mid-conversation](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#switching-models).
from line 1598
15791598 Cache prompt prefixes with `cache_control` to cut costs and latency, using automatic caching or explicit breakpoints with 5-minute or 1-hour TTLs.
15801599 </Card>
15811600</CardGroup>
1601
1602## Availability
1603
1604Preserved thinking is available on select models.
1605
1606### API support
1607
1608#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
1609
1610* Availability: Always on
1611* Usage: Keep requests append-only: send earlier turns back exactly as returned, and do not change `system`, `tools` or earlier messages.
1612
1613| Status | Cloud platforms | Models |
1614| --------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
1615| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview) |
1616
1617#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
1618
1619* Availability: Always on
1620* Usage: Managed Agents builds the requests.
1621
1622| Status | Cloud platforms | Models |
1623| --------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
1624| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview) |
15821625
build-with-claude/prompt-caching Changed · +72 / -6 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling
from line 2
22title: Prompt caching
33url: https://platform.claude.com/docs/en/build-with-claude/prompt-caching
44description: Cache prompt prefixes with `cache_control` to cut costs and latency, using automatic caching or explicit breakpoints with 5-minute or 1-hour TTLs.
5featureMetadata:
6 zdr:
7 eligibility: eligible
8 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). Your prompts and Claude's outputs are not stored. KV cache representations and cryptographic hashes are held in memory for the cache TTL and promptly deleted after expiry.
9 supportedModels:
10 - claude-fable-5-1
11 - claude-mythos-5-1
12 - claude-fable-5
13 - claude-mythos-5
14 - claude-mythos-preview
15 - claude-opus-5-5
16 - claude-opus-5
17 - claude-sonnet-5-5
18 - claude-sonnet-5
19 - claude-haiku-5-5
20 - claude-haiku-4-5-20251001
21 - claude-opus-4-8
22 - claude-opus-4-7
23 - claude-opus-4-6
24 - claude-sonnet-4-6
25 - claude-opus-4-5-20251101
26 - claude-sonnet-4-5-20250929
27 supportedPlatforms:
28 Claude API: ga
29 Claude Platform on AWS: ga
30 Claude Models on Google Cloud (Anthropic managed): ga
31 Amazon Bedrock: ga
32 Amazon Bedrock (InvokeModel): ga
33 Claude Models on Google Cloud (Google managed): ga
34 Microsoft Foundry: ga
535---
636
7Prompt caching optimizes your API usage by allowing resuming from specific prefixes in your prompts. This significantly reduces processing time and costs for repetitive tasks or prompts with consistent elements.
37**API support**: Usage differs by product.
838
9<Note>
10 To learn how zero data retention (ZDR) applies to this feature, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
11</Note>
39* **Messages API**: Opt in with `cache_control`, at the top level or on blocks.
40* **Managed Agents**: Always on. Cache entries use a 5-minute lifetime by default, and cache reads and writes show up in the session’s usage.
1241
42Prompt caching optimizes your API usage by allowing resuming from specific prefixes in your prompts. This significantly reduces processing time and costs for repetitive tasks or prompts with consistent elements.
43
1344There are two ways to enable prompt caching:
1445
1546* **[Automatic caching](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#automatic-caching):** Add a single `cache_control` field at the top level of your request. The system automatically applies the cache breakpoint to the last cacheable block and moves it forward as conversations grow. Best for multi-turn conversations where the growing message history should be cached automatically.
from line 3193
31623193
31633194KV (key-value) cache representations and cryptographic hashes of cached content are held in memory only and are not stored at rest. Cached entries have a minimum lifetime of 5 minutes (standard) or 1 hour (extended), after which they are promptly, though not immediately, deleted. Cache entries are isolated between organizations and, on the Claude API, Claude Platform on AWS, and Microsoft Foundry, between workspaces within an organization.
31643195
3165For ZDR eligibility across all features, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
3166
31673196***
31683197
31693198## FAQ
from line 3355
33263355 ```
33273356 </Accordion>
33283357</AccordionGroup>
3358
3359## Availability
3360
3361Prompt caching is available or in beta, depending on the API, and eligible for zero data retention and HIPAA readiness.
3362
3363### API support
3364
3365#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
3366
3367* Availability: Available
3368* Usage: Opt in with `cache_control`, at the top level or on blocks.
3369
3370| Status | Cloud platforms | Models |
3371| --------- | ----------------------- | ----------------------------------------------------------------- |
3372| Available | All supported platforms | [All models](https://platform.claude.com/docs/en/models/overview) |
3373
3374#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
3375
3376* Availability: Always on
3377* Usage: Cache entries use a 5-minute lifetime by default, and cache reads and writes show up in the session’s usage.
3378
3379| Status | Cloud platforms | Models |
3380| ------ | ----------------------- | ----------------------------------------------------------------- |
3381| Beta | All supported platforms | [All models](https://platform.claude.com/docs/en/models/overview) |
3382
3383### Data handling
3384
3385[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
3386
3387| Item | Zero data retention | HIPAA |
3388| ------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
3389| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
3390| Details | Your prompts and Claude’s outputs are not stored. KV cache representations and cryptographic hashes are held in memory for the cache TTL and promptly deleted after expiry. [Learn more](https://platform.claude.com/docs/en/build-with-claude/prompt-caching#data-retention) | — |
3391
3392Applicable endpoints:
3393
3394* `/v1/messages`
33293395
build-with-claude/refusals-and-fallback Changed · +38 / -0 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
from line 2
22title: Refusals and fallback
33url: https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback
44description: How Claude Fable models, Claude Opus models, Claude Sonnet 5.5, and Claude Haiku 5.5 return classifier refusals and how to retry refused requests on a fallback model.
5featureMetadata:
6 supportedModels:
7 - claude-fable-5-1
8 - claude-mythos-5-1
9 - claude-fable-5
10 - claude-opus-5-5
11 - claude-opus-5
12 - claude-sonnet-5-5
13 - claude-haiku-5-5
14 supportedPlatforms:
15 Claude API: ga
16 Claude Platform on AWS: ga
17 Claude Models on Google Cloud (Anthropic managed): ga
18 Amazon Bedrock: ga
19 Claude Models on Google Cloud (Google managed): ga
20 Microsoft Foundry: ga
521---
622
723Claude Fable 5.1, Claude Fable 5, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5.5, and Claude Haiku 5.5 include safety classifiers that can decline a request. When that happens, you receive a normal response, not an error, with `stop_reason: "refusal"`. Its `stop_details.category` names the policy area (see [What a refusal looks like](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#refusal-response)). You can usually still get an answer by sending the same request to another Claude model. This page shows you how to recognize a refusal and how to set up that retry.
from line 1194
11781194 Move an existing application to Claude Fable 5.1.
11791195 </Card>
11801196</CardGroup>
1197
1198## Availability
1199
1200Classifier refusals are available or in beta on select models, depending on the API.
1201
1202### API support
1203
1204#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
1205
1206* Availability: Always on
1207
1208| Status | Models |
1209| --------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
1210| Available | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview) |
1211
1212#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
1213
1214* Availability: Always on
1215
1216| Status | Cloud platforms | Models |
1217| ------ | ------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
1218| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview) |
11811219
build-with-claude/search-results Changed · +70 / -3 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling
from line 2
22title: Search results
33url: https://platform.claude.com/docs/en/build-with-claude/search-results
44description: Get natural citations for RAG applications by providing search results with source attribution
5featureMetadata:
6 status: ga
7 zdr:
8 eligibility: eligible
9 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
10 supportedModels:
11 - claude-fable-5-1
12 - claude-mythos-5-1
13 - claude-fable-5
14 - claude-mythos-5
15 - claude-mythos-preview
16 - claude-opus-5-5
17 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
21 - claude-haiku-4-5-20251001
22 - claude-opus-4-8
23 - claude-opus-4-7
24 - claude-opus-4-6
25 - claude-sonnet-4-6
26 - claude-opus-4-5-20251101
27 - claude-sonnet-4-5-20250929
28 supportedPlatforms:
29 Claude API: ga
30 Claude Platform on AWS: ga
31 Amazon Bedrock: ga
32 Amazon Bedrock (InvokeModel): ga
33 Claude Models on Google Cloud (Google managed): ga
34 Microsoft Foundry: ga
535---
636
7<Note>
8 To learn how zero data retention (ZDR) applies to this feature, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
9</Note>
37**API support**: Usage differs by product.
1038
39* **Messages API**: Always on. Send `search_result` blocks in a tool result or directly in a user message.
40* **Managed Agents**: Always on. Accepted only in tool results, such as a `user.custom_tool_result` event, not in `user.message` events.
41
1142Search result content blocks let Claude cite your own content the same way it cites web search results: each citation carries the source and title you provided. Use them in RAG (Retrieval-Augmented Generation) applications where Claude needs to attribute answers to your documents.
1243
1344All [active models](https://platform.claude.com/docs/en/models/overview) support search results with citations, with the exception of Claude Haiku 3. No beta header is required: search results are part of the standard Messages API.
from line 2187
21562187 Cache search results with `cache_control` to reduce cost and latency on repeated requests.
21572188 </Card>
21582189</CardGroup>
2190
2191## Availability
2192
2193Search results are available on every model and eligible for zero data retention and HIPAA readiness.
2194
2195### API support
2196
2197#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
2198
2199* Availability: Always on
2200* Usage: Send `search_result` blocks in a tool result or directly in a user message.
2201
2202| Status | Cloud platforms |
2203| --------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
2204| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
2205
2206#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
2207
2208* Availability: Always on
2209* Usage: Accepted only in tool results, such as a `user.custom_tool_result` event, not in `user.message` events.
2210
2211| Status | Cloud platforms |
2212| --------- | ------------------------------------------------------------------------------------------------------------------ |
2213| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) |
2214
2215### Data handling
2216
2217[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
2218
2219| Item | Zero data retention | HIPAA |
2220| ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
2221| Status | Eligible (excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
2222
2223Applicable endpoints:
2224
2225* `/v1/messages` with `search_results` source
21592226
build-with-claude/streaming Changed · +47 / -0 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
from line 2
22title: Streaming messages
33url: https://platform.claude.com/docs/en/build-with-claude/streaming
44description: Stream Messages API responses incrementally with server-sent events, including text, tool use, and thinking deltas.
5featureMetadata:
6 status: ga
7 supportedModels:
8 - claude-fable-5-1
9 - claude-mythos-5-1
10 - claude-fable-5
11 - claude-mythos-5
12 - claude-mythos-preview
13 - claude-opus-5-5
14 - claude-opus-5
15 - claude-sonnet-5-5
16 - claude-sonnet-5
17 - claude-haiku-5-5
18 - claude-haiku-4-5-20251001
19 - claude-opus-4-8
20 - claude-opus-4-7
21 - claude-opus-4-6
22 - claude-sonnet-4-6
23 - claude-opus-4-5-20251101
24 - claude-sonnet-4-5-20250929
25 supportedPlatforms:
26 Claude API: ga
27 Claude Platform on AWS: ga
28 Claude Models on Google Cloud (Anthropic managed): ga
29 Amazon Bedrock: ga
30 Amazon Bedrock (InvokeModel):
31 availability: ga
32 note: The older Amazon Bedrock InvokeModel integration does not stream plain server-sent events.
33 Claude Models on Google Cloud (Google managed): ga
34 Microsoft Foundry: ga
535---
636
737When creating a Message, you can set `"stream": true` to incrementally stream the response using [server-sent events](https://developer.mozilla.org/en-US/Web/API/Server-sent%5Fevents/Using%5Fserver-sent%5Fevents) (SSE).
from line 1560
15301560 Process large volumes of requests asynchronously when you don’t need real-time responses.
15311561 </Card>
15321562</CardGroup>
1563
1564## Availability
1565
1566Message streaming is available on every model.
1567
1568### API support
1569
1570#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
1571
1572* Availability: Available
1573
1574| Status | Cloud platforms |
1575| --------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
1576| Available | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Anthropic managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/anthropic-managed), [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) |
1577| Available<sup>1</sup> | [Amazon Bedrock (InvokeModel)](https://platform.claude.com/docs/en/build-with-claude/claude-on-amazon-bedrock-legacy) |
1578
15791. The older Amazon Bedrock InvokeModel integration does not stream plain server-sent events.
15331580
build-with-claude/structured-outputs Changed · +44 / -15 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Data handling ## Data retention
from line 5
55featureMetadata:
66 status: ga
77 zdr:
8 eligibility: eligible
9 note: Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements).
8 eligibility: qualified
9 note: "Excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements). Your prompts and Claude's outputs are not stored. Only the JSON schema is cached, for up to 24 hours since last use. This also covers strict tool use (`strict: true` on tools), which uses the same grammar pipeline."
1010 supportedModels:
1111 - claude-fable-5-1
1212 - claude-mythos-5-1
from line 15
1515 - claude-mythos-preview
1616 - claude-opus-5-5
1717 - claude-opus-5
18 - claude-sonnet-5-5
19 - claude-sonnet-5
20 - claude-haiku-5-5
21 - claude-haiku-4-5-20251001
1822 - claude-opus-4-8
1923 - claude-opus-4-7
2024 - claude-opus-4-6
21 - claude-sonnet-5-5
22 - claude-sonnet-5
2325 - claude-sonnet-4-6
24 - claude-sonnet-4-5-20250929
2526 - claude-opus-4-5-20251101
26 - claude-haiku-5-5
27 - claude-haiku-4-5-20251001
27 - claude-sonnet-4-5-20250929
2828 supportedPlatforms:
2929 Claude API: ga
3030 Claude Platform on AWS: ga
31 Claude Models on Google Cloud (Anthropic managed): ga
3132 Amazon Bedrock: ga
33 Amazon Bedrock (InvokeModel): ga
3234 Claude Models on Google Cloud (Google managed): ga
3335 Microsoft Foundry: ga
3436---
from line 2787
27852787
27862788The Python SDK (v1.0 and later) does not accept `output_format={...}` on `client.beta.messages.create()` or `count_tokens()` and raises a `TypeError`. Use `output_config` instead. See [Using a raw JSON schema](https://platform.claude.com/docs/en/build-with-claude/structured-outputs#using-a-raw-json-schema) for the updated API shape.
27872789
2788## Data retention
2789
2790Prompts and responses are processed with ZDR when using structured outputs. However, the JSON schema itself is temporarily cached for up to 24 hours since last use for optimization purposes. No prompt or response data is retained beyond the API response.
2791
2792Structured outputs are HIPAA eligible, but **PHI must not be included in JSON schema definitions**. The API compiles JSON schemas into grammars that are cached separately from message content, and these cached schemas do not receive the same PHI protections as prompts and responses. Do not include PHI in schema property names, `enum` values, `const` values, or `pattern` regular expressions. PHI should only appear in message content (prompts and responses), where it is protected under HIPAA safeguards.
2793
2794For ZDR and HIPAA eligibility across all features, see [API and data retention](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention).
2795
27962790## Feature compatibility
27972791
27982792**Works with:**
from line 2804
28102804<Tip>
28112805 **Grammar scope:** Grammars apply only to Claude’s direct output, not to tool use calls, tool results, or thinking tags (when using [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking)). Grammar state resets between sections, allowing Claude to think freely while still producing structured output in the final response.
28122806</Tip>
2807
2808## Availability
2809
2810Structured outputs are available on every model and eligible for HIPAA readiness.
2811
2812### API support
2813
2814#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
2815
2816* Availability: Available
2817
2818| Status | Cloud platforms | Models |
2819| --------- | ----------------------- | ----------------------------------------------------------------- |
2820| Available | All supported platforms | [All models](https://platform.claude.com/docs/en/models/overview) |
2821
2822#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
2823
2824* Availability: Not available
2825
2826### Data handling
2827
2828[Data handling overview](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#feature-eligibility)
2829
2830| Item | Zero data retention | HIPAA |
2831| ------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
2832| Status | Eligible (qualified; excludes [Covered Models](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#model-specific-data-retention-requirements)) | Eligible |
2833| Details | Your prompts and Claude’s outputs are not stored. Only the JSON schema is cached, for up to 24 hours since last use. This also covers strict tool use (`strict: true` on tools), which uses the same grammar pipeline. [Learn more](https://platform.claude.com/docs/en/build-with-claude/structured-outputs#data-retention) | PHI must not be included in JSON schema definitions; see PHI handling guidelines. [Learn more](https://platform.claude.com/docs/en/manage-claude/api-and-data-retention#phi-handling-guidelines) |
2834
2835Applicable endpoints:
2836
2837* `/v1/messages`
2838
2839Prompts and responses are processed with ZDR when using structured outputs. However, the JSON schema itself is temporarily cached for up to 24 hours since last use for optimization purposes. No prompt or response data is retained beyond the API response.
2840
2841Structured outputs are HIPAA eligible, but **PHI must not be included in JSON schema definitions**. The API compiles JSON schemas into grammars that are cached separately from message content, and these cached schemas do not receive the same PHI protections as prompts and responses. Do not include PHI in schema property names, `enum` values, `const` values, or `pattern` regular expressions. PHI should only appear in message content (prompts and responses), where it is protected under HIPAA safeguards.
28132842
28142843## Next steps
28152844
build-with-claude/task-budgets Changed · +36 / -2 lines
## Availability ### API support #### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages) #### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) ### Required beta headers
from line 10
1010 - claude-mythos-5-1
1111 - claude-fable-5
1212 - claude-mythos-5
13 - claude-mythos-preview
1314 - claude-opus-5-5
1415 - claude-opus-5
15 - claude-opus-4-8
16 - claude-opus-4-7
1716 - claude-sonnet-5-5
1817 - claude-haiku-5-5
18 - claude-opus-4-8
19 - claude-opus-4-7
20 supportedPlatforms:
21 Claude API: beta
22 Claude Platform on AWS: beta
23 Claude Models on Google Cloud (Google managed): beta
24 Microsoft Foundry: beta
1925---
2026
2127Task budgets let you tell Claude how many tokens it has for a full agentic loop, including thinking, tool calls, tool results, and output. The model receives a running countdown and uses it to prioritize work and finish gracefully as the budget is consumed.
from line 702
696702 Reduce cost and latency on repeated prompts by caching prompt prefixes.
697703 </Card>
698704</CardGroup>
705
706## Availability
707
708Task budgets are in beta on select models.
709
710### API support
711
712#### [Messages API](https://platform.claude.com/docs/en/build-with-claude/working-with-messages)
713
714* Availability: Available
715* Usage: Advisory: set in `output_config.task_budget`. The API does not stop Claude at the budget.
716
717| Status | Cloud platforms | Models |
718| ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
719| Beta | Claude API, [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude Models on Google Cloud (Google managed)](https://platform.claude.com/docs/en/build-with-claude/claude-models-on-google-cloud/google-managed), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview), [Claude Mythos 5.1](https://platform.claude.com/docs/en/models/mythos-5-1/overview), [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [Claude Mythos 5](https://platform.claude.com/docs/en/models/mythos-5/overview), Claude Mythos Preview, [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview), [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview), [Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/overview), [Claude Haiku 5.5](https://platform.claude.com/docs/en/models/haiku-5-5/overview), [Claude Opus 4.8](https://platform.claude.com/docs/en/models/opus-4-8/overview), [Claude Opus 4.7](https://platform.claude.com/docs/en/models/opus-4-7/overview) |
720
721#### [Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview)
722
723* Availability: Not available
724* Usage: Use [Session budgets](https://platform.claude.com/docs/en/managed-agents/budgets) instead.
725
726### Required beta headers
727
728[Beta headers overview](https://platform.claude.com/docs/en/api/beta-headers)
729
730Beta headers allow you to access experimental features and new model capabilities before they become part of the standard API.
731
732* `task-budgets-2026-03-13`
699733