Follow Discord
Sweep 22 Sep 2026 · 17:19Z Build v2.1.280 501 read Stable v2.1.267 Latest v2.1.280 Next v2.1.280 Feeds RSS JSON llms.txt Unofficial
Home Docs Claude Developer Platform
One capture · api

One read of Claude Developer Platform

147 pages moved out of 634 read.

api-20260922T163722Z

Pages moved 147 significant first
Pages read 634 in this capture
Captured 16:37 UTC
Corpus hash 4b35df8b3fd5 corpus-hash

What this read moved

101–125 of 147

This capture is too large to show at once. Changes 101-125 of 147 are below, significant first; the rest are on the following screens.

models/fable-5-1/overview Changed · +3 / -3 lines

from line 31
3131| Model | Context | Max output | Price / MTok | Latency | Thinking | Default effort | Knowledge cutoff |
3232| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------- | :------------------- | :------------- | :--------------- |
3333| **Claude Fable 5.1** (this model) | 1M | 128K | $10 / $50 | Slower | Adaptive (always on) | `high` | Jun 2026 |
34| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Moderate | Adaptive | `high` | May 2026 |
34| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Moderate | Adaptive (always on) | `medium` | Jun 2026 |
3535| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Fast | Adaptive | `high` | Jan 2026 |
3636| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Fastest | Extended | — | Feb 2025 |
3737 
3838* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
39* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
40* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
39* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
40* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
4141* **Latency:** Comparative latency, relative to the current lineup, as published in the models overview. Actual latency depends on prompt length, output length, and thinking effort.
4242* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
4343* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.

models/fable-5/overview Changed · +3 / -3 lines

from line 24
2424| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------------------- | :------------- | :--------------- |
2525| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
2626| **Claude Fable 5** (this model) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jan 2026 |
27| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
27| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2828| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2929| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
3030 
3131* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
32* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
33* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
32* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
33* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3434* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3535* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3636* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.

models/haiku-4-5/overview Changed · +3 / -3 lines

from line 19
1919| Model | Context | Max output | Price / MTok | Latency | Thinking | Default effort | Knowledge cutoff |
2020| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------- | :------------------- | :------------- | :--------------- |
2121| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Slower | Adaptive (always on) | `high` | Jun 2026 |
22| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Moderate | Adaptive | `high` | May 2026 |
22| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Moderate | Adaptive (always on) | `medium` | Jun 2026 |
2323| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Fast | Adaptive | `high` | Jan 2026 |
2424| **Claude Haiku 4.5** (this model) | 200K | 64K | $1 / $5 | Fastest | Extended | — | Feb 2025 |
2525 
2626* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
27* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
28* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
27* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
28* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
2929* **Latency:** Comparative latency, relative to the current lineup, as published in the models overview. Actual latency depends on prompt length, output length, and thinking effort.
3030* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3131* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.

models/mythos-5-1/overview Changed · +3 / -3 lines

from line 22
2222| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------- | :------------------- | :------------- | :--------------- |
2323| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Slower | Adaptive (always on) | `high` | Jun 2026 |
2424| **Claude Mythos 5.1** (this model) | 1M | 128K | $10 / $50 | Slower | Adaptive (always on) | `high` | Jun 2026 |
25| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Moderate | Adaptive | `high` | May 2026 |
25| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Moderate | Adaptive (always on) | `medium` | Jun 2026 |
2626| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Fast | Adaptive | `high` | Jan 2026 |
2727| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Fastest | Extended | — | Feb 2025 |
2828 
2929* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
30* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
31* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
30* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
31* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3232* **Latency:** Comparative latency, relative to the current lineup, as published in the models overview. Actual latency depends on prompt length, output length, and thinking effort.
3333* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3434* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.

models/mythos-5/overview Changed · +3 / -3 lines

from line 22
2222| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------------------- | :------------- | :--------------- |
2323| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
2424| **Claude Mythos 5** (this model) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jan 2026 |
25| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
25| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2626| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2727| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
2828 
2929* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
30* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
31* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
30* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
31* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3232* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3333* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3434* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.

models/opus-4-5/overview Changed · +8 / -8 lines

from line 1
11---
22title: Claude Opus 4.5
33url: https://platform.claude.com/docs/en/models/opus-4-5/overview
4description: "Claude Opus 4.5 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.5 is a legacy model; Claude Opus 5 is the current Opus model."
4description: "Claude Opus 4.5 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.5 is a legacy model; Claude Opus 5.5 is the current Opus model."
55---
66 
77**Legacy.** Released November 24, 2025.
88 
9Although Claude Opus 4.5 is still available, you should consider migrating to Claude Opus 5 for improved performance. [See Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) · [Migrate to Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-45)
9Although Claude Opus 4.5 is still available, you should consider migrating to Claude Opus 5.5 for improved performance. [See Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) · [Migrate to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide)
1010 
1111Model ID: `claude-opus-4-5-20251101`
1212 
from line 19
1919| Model | Context | Max output | Price / MTok | Thinking | Default effort | Knowledge cutoff |
2020| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------------------- | :------------- | :--------------- |
2121| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
22| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
22| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2323| **Claude Opus 4.5** (this model) | 200K | 64K | $5 / $25 | Extended | `high` | May 2025 |
2424| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2525| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
2626 
2727* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3030* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3131* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3232* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.
from line 80
8080## Resources
8181 
8282<CardGroup cols={3}>
83 <Card title="Migrate to Claude Opus 5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-45">
84 What changes when moving from Claude Opus 4.6 and earlier Opus models to Claude Opus 5.
83 <Card title="Migrate to Claude Opus 5.5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-47">
84 What changes when moving from Claude Opus 4.7 and earlier Opus models to Claude Opus 5.5.
8585 </Card>
8686 
87 <Card title="Claude Opus 5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5/overview">
87 <Card title="Claude Opus 5.5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5-5/overview">
8888 The current Opus model: overview, specs, and resources.
8989 </Card>
9090</CardGroup>

models/opus-4-6/overview Changed · +8 / -8 lines

from line 1
11---
22title: Claude Opus 4.6
33url: https://platform.claude.com/docs/en/models/opus-4-6/overview
4description: "Claude Opus 4.6 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.6 is a legacy model; Claude Opus 5 is the current Opus model."
4description: "Claude Opus 4.6 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.6 is a legacy model; Claude Opus 5.5 is the current Opus model."
55---
66 
77**Legacy.** Released February 5, 2026.
88 
9Although Claude Opus 4.6 is still available, you should consider migrating to Claude Opus 5 for improved performance. [See Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) · [Migrate to Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-46)
9Although Claude Opus 4.6 is still available, you should consider migrating to Claude Opus 5.5 for improved performance. [See Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) · [Migrate to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide)
1010 
1111Model ID: `claude-opus-4-6`
1212 
from line 19
1919| Model | Context | Max output | Price / MTok | Thinking | Default effort | Knowledge cutoff |
2020| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :----------------------------- | :------------- | :--------------- |
2121| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
22| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
22| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2323| **Claude Opus 4.6** (this model) | 1M | 128K | $5 / $25 | Adaptive (extended deprecated) | `high` | May 2025 |
2424| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2525| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
2626 
2727* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3030* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3131* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3232* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.
from line 80
8080## Resources
8181 
8282<CardGroup cols={3}>
83 <Card title="Migrate to Claude Opus 5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-46">
84 What changes when moving from Claude Opus 4.6 and earlier Opus models to Claude Opus 5.
83 <Card title="Migrate to Claude Opus 5.5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-47">
84 What changes when moving from Claude Opus 4.7 and earlier Opus models to Claude Opus 5.5.
8585 </Card>
8686 
87 <Card title="Claude Opus 5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5/overview">
87 <Card title="Claude Opus 5.5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5-5/overview">
8888 The current Opus model: overview, specs, and resources.
8989 </Card>
9090</CardGroup>

models/opus-4-7/overview Changed · +8 / -8 lines

from line 1
11---
22title: Claude Opus 4.7
33url: https://platform.claude.com/docs/en/models/opus-4-7/overview
4description: "Claude Opus 4.7 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.7 is a legacy model; Claude Opus 5 is the current Opus model."
4description: "Claude Opus 4.7 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.7 is a legacy model; Claude Opus 5.5 is the current Opus model."
55---
66 
77**Legacy.** Released April 16, 2026.
88 
9Although Claude Opus 4.7 is still available, you should consider migrating to Claude Opus 5 for improved performance. [See Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) · [Migrate to Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-47)
9Although Claude Opus 4.7 is still available, you should consider migrating to Claude Opus 5.5 for improved performance. [See Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) · [Migrate to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide)
1010 
1111Model ID: `claude-opus-4-7`
1212 
from line 19
1919| Model | Context | Max output | Price / MTok | Thinking | Default effort | Knowledge cutoff |
2020| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------------------- | :------------- | :--------------- |
2121| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
22| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
22| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2323| **Claude Opus 4.7** (this model) | 1M | 128K | $5 / $25 | Adaptive | `high` | Jan 2026 |
2424| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2525| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
2626 
2727* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3030* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3131* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3232* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.
from line 80
8080## Resources
8181 
8282<CardGroup cols={3}>
83 <Card title="Migrate to Claude Opus 5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-47">
84 What changes when moving from Claude Opus 4.7 to Claude Opus 5.
83 <Card title="Migrate to Claude Opus 5.5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-47">
84 What changes when moving from Claude Opus 4.7 to Claude Opus 5.5.
8585 </Card>
8686 
87 <Card title="Claude Opus 5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5/overview">
87 <Card title="Claude Opus 5.5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5-5/overview">
8888 The current Opus model: overview, specs, and resources.
8989 </Card>
9090</CardGroup>

models/opus-4-8/overview Changed · +8 / -8 lines

from line 1
11---
22title: Claude Opus 4.8
33url: https://platform.claude.com/docs/en/models/opus-4-8/overview
4description: "Claude Opus 4.8 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.8 is a legacy model; Claude Opus 5 is the current Opus model."
4description: "Claude Opus 4.8 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 4.8 is a legacy model; Claude Opus 5.5 is the current Opus model."
55---
66 
77**Legacy.** Released May 28, 2026.
88 
9Although Claude Opus 4.8 is still available, you should consider migrating to Claude Opus 5 for improved performance. [See Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) · [Migrate to Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-4-8-to-claude-opus-5)
9Although Claude Opus 4.8 is still available, you should consider migrating to Claude Opus 5.5 for improved performance. [See Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) · [Migrate to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide)
1010 
1111Model ID: `claude-opus-4-8`
1212 
from line 17
1717| Model | Context | Max output | Price / MTok | Thinking | Default effort | Knowledge cutoff |
1818| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------------------- | :------------- | :--------------- |
1919| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
20| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
20| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2121| **Claude Opus 4.8** (this model) | 1M | 128K | $5 / $25 | Adaptive | `high` | Jan 2026 |
2222| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2323| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
2424 
2525* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
26* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
27* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
26* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
27* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
2828* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
2929* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3030* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.
from line 78
7878## Resources
7979 
8080<CardGroup cols={3}>
81 <Card title="Migrate to Claude Opus 5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5/migration-guide#migrating-from-claude-opus-4-8-to-claude-opus-5">
82 What changes when moving from Claude Opus 4.8 to Claude Opus 5.
81 <Card title="Migrate to Claude Opus 5.5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-4-8">
82 What changes when moving from Claude Opus 4.8 to Claude Opus 5.5.
8383 </Card>
8484 
85 <Card title="Claude Opus 5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5/overview">
85 <Card title="Claude Opus 5.5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5-5/overview">
8686 The current Opus model: overview, specs, and resources.
8787 </Card>
8888 

models/opus-5/overview Changed · +23 / -14 lines

from line 1
11---
22title: Claude Opus 5
33url: https://platform.claude.com/docs/en/models/opus-5/overview
4description: "Claude Opus 5 at a glance: what it's for, model IDs on every platform, context window, output limits, pricing, availability, and the guides and resources for building with it."
4description: "Claude Opus 5 reference: lifecycle status, model IDs on every platform, context window, output limits, pricing, and migration resources. Claude Opus 5.5 is the current Opus model."
55---
66 
7**Latest.** Released July 24, 2026.
7**Legacy.** Released July 24, 2026.
88 
99For complex agentic coding and enterprise work
1010 
11Although Claude Opus 5 is still available, you should consider migrating to Claude Opus 5.5 for improved performance. [See Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) · [Migrate to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide)
12 
1113Model ID: `claude-opus-5`
1214 
1315Context window: 1M tokens · Max output: 128K tokens · Input pricing: $5 / MTok · Output pricing: $25 / MTok
1416 
15[Announcement](https://www.anthropic.com/news/claude-opus-5) · [What’s new](https://platform.claude.com/docs/en/models/opus-5/whats-new-opus-5) · [Migration guide](https://platform.claude.com/docs/en/models/opus-5/migration-guide)
17[Announcement](https://www.anthropic.com/news/claude-opus-5) · [What’s new](https://platform.claude.com/docs/en/models/opus-5/whats-new-opus-5)
1618 
1719## Overview
1820 
from line 24
2224 
2325## How it compares
2426 
25| Model | Context | Max output | Price / MTok | Latency | Thinking | Default effort | Knowledge cutoff |
26| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------- | :------------------- | :------------- | :--------------- |
27| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Slower | Adaptive (always on) | `high` | Jun 2026 |
28| **Claude Opus 5** (this model) | 1M | 128K | $5 / $25 | Moderate | Adaptive | `high` | May 2026 |
29| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Fast | Adaptive | `high` | Jan 2026 |
30| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Fastest | Extended | — | Feb 2025 |
27| Model | Context | Max output | Price / MTok | Thinking | Default effort | Knowledge cutoff |
28| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------------------- | :------------- | :--------------- |
29| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
30| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
31| **Claude Opus 5** (this model) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
32| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
33| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
3134 
3235* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
33* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
34* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
35* **Latency:** Comparative latency, relative to the current lineup, as published in the models overview. Actual latency depends on prompt length, output length, and thinking effort.
36* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
37* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3638* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3739* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3840* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.
from line 72
7072| [Max output (Batch API, beta)](https://platform.claude.com/docs/en/build-with-claude/batch-processing#extended-output-beta) | 300K tokens |
7173| [Thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) | Adaptive |
7274| [Default effort](https://platform.claude.com/docs/en/build-with-claude/effort) | `high` |
73| Comparative latency | Moderate |
7475| Input → output | Text and images → text |
7576| Reliable knowledge cutoff | May 2026 |
7677| Training data cutoff | May 2026 |
from line 80
7980 
8081| Feature | Value |
8182| :---------------------------------------------------------------------------- | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
82| [Status](https://platform.claude.com/docs/en/about-claude/model-deprecations) | Active (latest) |
83| [Status](https://platform.claude.com/docs/en/about-claude/model-deprecations) | Active (legacy) |
8384| Released | July 24, 2026 |
8485| Retirement | Not sooner than July 24, 2027 |
8586| Platforms | Claude API, [Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Google Cloud](https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai), [Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry), [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) |
from line 94
9394## Resources
9495 
9596<CardGroup cols={3}>
97 <Card title="Migrate to Claude Opus 5.5" icon="arrows-left-right" href="https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-5">
98 What changes when moving from Claude Opus 5 to Claude Opus 5.5.
99 </Card>
100 
101 <Card title="Claude Opus 5.5" icon="arrow-right" href="https://platform.claude.com/docs/en/models/opus-5-5/overview">
102 The current Opus model: overview, specs, and resources.
103 </Card>
104 
96105 <Card title="Prompting Claude Opus 5" icon="lightbulb" href="https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5">
97106 Model-specific prompting guidance.
98107 </Card>

models/opus-5/whats-new-opus-5 Changed · +5 / -1 lines

from line 4
44description: Overview of new features and behavior changes in Claude Opus 5.
55---
66 
7<Note>
8 Claude Opus 5.5 is the current Opus model. See [What's new in Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/whats-new-opus-5-5).
9</Note>
10 
711Claude Opus 5 is a step-change improvement over Claude Opus 4.8, with the largest gains in deep reasoning, agentic and long-horizon tasks, and test-time compute scaling. This page summarizes everything new in Claude Opus 5, including mid-conversation tool changes and two breaking changes for code running on Claude Opus 4.8: thinking is on by default, and thinking can be disabled only at effort `high` or below.
812 
913## New model
from line 240
236240 
237241### Disabling thinking requires effort `high` or below
238242 
239On Claude Opus 5, `thinking: {"type": "disabled"}` is accepted only when the effort level is `high` or below. Setting `thinking: {"type": "disabled"}` with effort `xhigh` or `max` returns a 400 error. This rule is enforced on every request to Claude Opus 5 and later models. It is a breaking change from Claude Opus 4.8, where disabling thinking was independent of the effort level. If your Claude Opus 4.8 requests disable thinking at effort `xhigh` or `max`, either keep thinking disabled and set effort to `high` or below, or keep the effort level and remove the `thinking` field.
243On Claude Opus 5, `thinking: {"type": "disabled"}` is accepted only when the effort level is `high` or below. Setting `thinking: {"type": "disabled"}` with effort `xhigh` or `max` returns a 400 error. This rule is enforced on every request to Claude Opus 5. It is a breaking change from Claude Opus 4.8, where disabling thinking was independent of the effort level. If your Claude Opus 4.8 requests disable thinking at effort `xhigh` or `max`, either keep thinking disabled and set effort to `high` or below, or keep the effort level and remove the `thinking` field.
240244 
241245With thinking disabled, Claude Opus 5 can occasionally write a tool call into its text output instead of emitting a `tool_use` block, or include internal XML tags in its visible response. Where possible, keep thinking enabled and control token cost with lower effort levels; for integrations that must keep thinking disabled, see [Running with thinking disabled](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5#running-with-thinking-disabled) for prompting mitigations.
242246 

models/overview Changed · +24 / -24 lines

from line 22
2222 
2323## Compare models
2424 
25If you're unsure which model to use, start with [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) for most workloads. Use [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) for demanding reasoning and long-horizon agentic work, or when your evals on Claude Opus 5 at higher effort still fall short. All current models support text and image input, text output, multilingual capabilities, vision, and tool use. Each model's page lists the platforms it's available on.
25If you're unsure which model to use, start with [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) for most workloads. Use [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) for demanding reasoning and long-horizon agentic work, or when your evals on Claude Opus 5.5 at higher effort still fall short. All current models support text and image input, text output, multilingual capabilities, vision, and tool use. Each model's page lists the platforms it's available on.
2626 
27| Feature | Claude Fable 5.1 | Claude Opus 5 | Claude Sonnet 5 | Claude Haiku 4.5 |
28| :-------------------------------------------------------------------------------------------------------- | :-------------------------------------------------------------------------------- | :-------------------------------------------------------------------------- | :------------------------------------------------------------------------------ | :-------------------------------------------------------------------------------- |
29| Description | For demanding reasoning and long-horizon agentic work | For complex agentic coding and enterprise work | The best combination of speed and intelligence | The fastest model with near-frontier intelligence |
30| Model page | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) |
31| Comparative latency | Slower | Moderate | Fast | Fastest |
32| [Pricing](https://platform.claude.com/docs/en/about-claude/pricing) | $10 / input MTok, $50 / output MTok | $5 / input MTok, $25 / output MTok | $2 / input MTok, $10 / output MTok | $1 / input MTok, $5 / output MTok |
33| Claude API ID | `claude-fable-5-1` | `claude-opus-5` | `claude-sonnet-5` | `claude-haiku-4-5-20251001` |
34| [Thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) | Adaptive (always on) | Adaptive | Adaptive | Extended |
35| [Default effort](https://platform.claude.com/docs/en/build-with-claude/effort) | `high` | `high` | `high` | Not supported |
36| [Context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) | 1M tokens | 1M tokens | 1M tokens | 200K tokens |
37| Max output | 128K tokens | 128K tokens | 128K tokens | 64K tokens |
38| Reliable knowledge cutoff | Jun 2026 | May 2026 | Jan 2026 | Feb 2025 |
39| Training data cutoff | Jun 2026 | May 2026 | Jan 2026 | Jul 2025 |
40| [Retirement](https://platform.claude.com/docs/en/about-claude/model-deprecations) | Not sooner than September 1, 2027 | Not sooner than July 24, 2027 | Not sooner than June 30, 2027 | Not sooner than October 15, 2026 |
41| Claude API alias | `claude-fable-5-1` | `claude-opus-5` | `claude-sonnet-5` | `claude-haiku-4-5` |
42| [Amazon Bedrock ID](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock) | `anthropic.claude-fable-5-1` | `anthropic.claude-opus-5` | `anthropic.claude-sonnet-5` | `anthropic.claude-haiku-4-5` |
43| [Google Cloud ID](https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai) | `claude-fable-5-1` | `claude-opus-5` | `claude-sonnet-5` | `claude-haiku-4-5@20251001` |
44| [Microsoft Foundry ID](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | `claude-fable-5-1` | `claude-opus-5` | `claude-sonnet-5` | `claude-haiku-4-5` |
45| [Claude Platform on AWS ID](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) | `claude-fable-5-1` | `claude-opus-5` | `claude-sonnet-5` | `claude-haiku-4-5` |
27| Feature | Claude Fable 5.1 | Claude Opus 5.5 | Claude Sonnet 5 | Claude Haiku 4.5 |
28| :-------------------------------------------------------------------------------------------------------- | :-------------------------------------------------------------------------------- | :------------------------------------------------------------------------------ | :------------------------------------------------------------------------------ | :-------------------------------------------------------------------------------- |
29| Description | For demanding reasoning and long-horizon agentic work | For long-running agentic coding and knowledge work | The best combination of speed and intelligence | The fastest model with near-frontier intelligence |
30| Model page | [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) |
31| Comparative latency | Slower | Moderate | Fast | Fastest |
32| [Pricing](https://platform.claude.com/docs/en/about-claude/pricing) | $10 / input MTok, $50 / output MTok | $4 / input MTok, $20 / output MTok | $2 / input MTok, $10 / output MTok | $1 / input MTok, $5 / output MTok |
33| Claude API ID | `claude-fable-5-1` | `claude-opus-5-5` | `claude-sonnet-5` | `claude-haiku-4-5-20251001` |
34| [Thinking](https://platform.claude.com/docs/en/build-with-claude/thinking) | Adaptive (always on) | Adaptive (always on) | Adaptive | Extended |
35| [Default effort](https://platform.claude.com/docs/en/build-with-claude/effort) | `high` | `medium` | `high` | Not supported |
36| [Context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) | 1M tokens | 1M tokens | 1M tokens | 200K tokens |
37| Max output | 128K tokens | 128K tokens | 128K tokens | 64K tokens |
38| Reliable knowledge cutoff | Jun 2026 | Jun 2026 | Jan 2026 | Feb 2025 |
39| Training data cutoff | Jun 2026 | Jun 2026 | Jan 2026 | Jul 2025 |
40| [Retirement](https://platform.claude.com/docs/en/about-claude/model-deprecations) | Not sooner than September 1, 2027 | Not sooner than September 22, 2027 | Not sooner than June 30, 2027 | Not sooner than October 15, 2026 |
41| Claude API alias | `claude-fable-5-1` | `claude-opus-5-5` | `claude-sonnet-5` | `claude-haiku-4-5` |
42| [Amazon Bedrock ID](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock) | `anthropic.claude-fable-5-1` | `anthropic.claude-opus-5-5` | `anthropic.claude-sonnet-5` | `anthropic.claude-haiku-4-5` |
43| [Google Cloud ID](https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai) | `claude-fable-5-1` | `claude-opus-5-5` | `claude-sonnet-5` | `claude-haiku-4-5@20251001` |
44| [Microsoft Foundry ID](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry) | `claude-fable-5-1` | `claude-opus-5-5` | `claude-sonnet-5` | `claude-haiku-4-5` |
45| [Claude Platform on AWS ID](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws) | `claude-fable-5-1` | `claude-opus-5-5` | `claude-sonnet-5` | `claude-haiku-4-5` |
4646 
4747* **Comparative latency:** Relative to the current lineup. Actual latency depends on prompt length, output length, and thinking effort.
48* **Pricing:** Base price per million tokens. Batch API requests are 50% off; prompt cache reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for cache writes, long-context, and per-platform pricing.
48* **Pricing:** Base price per million tokens. Batch API requests are 50% off; prompt cache reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for cache writes, long-context, and per-platform pricing.
4949* **Claude API ID:** Every Claude model ID is a pinned snapshot, including the dateless IDs used from the 4.6 generation on.
5050* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual thinking.type “enabled” + budget\_tokens mode on earlier models; it is deprecated on Claude Opus 4.6 and Claude Sonnet 4.6 and not accepted on later models.
5151* **Default effort:** The effort parameter’s default on the Claude API. Set effort explicitly to use a different level.
5252* **Context window:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
53* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
53* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
5454* **Reliable knowledge cutoff:** The date through which the model’s knowledge is most extensive and reliable. Training data cutoff (under Show all details) is the broader range of data used. See Anthropic’s Transparency Hub for details.
5555* **Retirement:** Anthropic’s commitment for Anthropic-operated platforms (Claude API, Claude Platform on AWS, Microsoft Foundry). Amazon Bedrock and Google Cloud set their own dates.
5656* **Claude API alias:** For models before the 4.6 generation, the alias is a convenience pointer that resolves to the dated ID. Dateless IDs are their own pinned snapshot; the alias row repeats them.
from line 61
6161 
6262See [Model IDs and versioning](https://platform.claude.com/docs/en/about-claude/models/model-ids-and-versions) and [Pricing](https://platform.claude.com/docs/en/about-claude/pricing).
6363 
64Legacy models (still available): [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-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 Opus 4.5](https://platform.claude.com/docs/en/models/opus-4-5/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview), [Claude Sonnet 4.5](https://platform.claude.com/docs/en/models/sonnet-4-5/overview).
64Legacy models (still available): [Claude Fable 5](https://platform.claude.com/docs/en/models/fable-5/overview), [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), [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 Opus 4.5](https://platform.claude.com/docs/en/models/opus-4-5/overview), [Claude Sonnet 4.6](https://platform.claude.com/docs/en/models/sonnet-4-6/overview), [Claude Sonnet 4.5](https://platform.claude.com/docs/en/models/sonnet-4-5/overview).
6565 
6666Once you've picked a model, [learn how to make your first API call](https://platform.claude.com/docs/en/get-started). To understand how model IDs, aliases, and snapshots work, see [Model IDs and versioning](https://platform.claude.com/docs/en/about-claude/models/model-ids-and-versions); for the reliable-knowledge and training-data cutoffs behind each model, see [Anthropic's Transparency Hub](https://www.anthropic.com/transparency).
6767 
from line 75
7575 
7676* **Performance:** Top-tier results in reasoning, coding, multilingual tasks, long-context handling, honesty, and image processing. See [Prompting best practices](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/claude-prompting-best-practices) for general and model-specific prompting guidance.
7777* **Engaging responses:** Claude models are ideal for applications that require rich, human-like interactions. If you prefer more concise responses, adjust your prompts to guide the model toward the desired output length. Refer to the [prompt engineering guides](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering) for details.
78* **Output quality:** When migrating from a previous model generation, you may notice larger improvements in overall performance. If you're on Claude Opus 4.8 or earlier, see [Migrating to Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/migration-guide).
78* **Output quality:** When migrating from a previous model generation, you may notice larger improvements in overall performance. If you're on Claude Opus 5 or earlier, see [Migrating to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide).
7979 
8080## Get started with Claude
8181 

models/sonnet-4-5/overview Changed · +3 / -3 lines

from line 19
1919| Model | Context | Max output | Price / MTok | Thinking | Default effort | Knowledge cutoff |
2020| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------------------- | :------------- | :--------------- |
2121| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
22| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
22| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2323| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2424| **Claude Sonnet 4.5** (this model) | 200K | 64K | $3 / $15 | Extended | — | Jan 2025 |
2525| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
2626 
2727* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3030* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3131* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3232* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.

models/sonnet-4-6/overview Changed · +3 / -3 lines

from line 19
1919| Model | Context | Max output | Price / MTok | Thinking | Default effort | Knowledge cutoff |
2020| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :----------------------------- | :------------- | :--------------- |
2121| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Adaptive (always on) | `high` | Jun 2026 |
22| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Adaptive | `high` | May 2026 |
22| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Adaptive (always on) | `medium` | Jun 2026 |
2323| [Claude Sonnet 5](https://platform.claude.com/docs/en/models/sonnet-5/overview) | 1M | 128K | $2 / $10 | Adaptive | `high` | Jan 2026 |
2424| **Claude Sonnet 4.6** (this model) | 1M | 128K | $3 / $15 | Adaptive (extended deprecated) | `high` | Aug 2025 |
2525| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Extended | — | Feb 2025 |
2626 
2727* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
28* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
29* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3030* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3131* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.
3232* **Knowledge cutoff:** Reliable knowledge cutoff: the date through which the model’s knowledge is most extensive and reliable.

models/sonnet-5/migration-guide Changed · +9 / -9 lines

from line 75
7575 "messages": [
7676 {
7777 "role": "user",
78 "content": "Are there an infinite number of prime numbers such that n mod 4 == 3?"
78 "content": "Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024."
7979 }
8080 ]
8181 }'
from line 92
9292 effort: high
9393 messages:
9494 - role: user
95 content: Are there an infinite number of prime numbers such that n mod 4 == 3?
95 content: Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024.
9696 YAML
9797 ```
9898 
from line 107
107107 messages=[
108108 {
109109 "role": "user",
110 "content": "Are there an infinite number of prime numbers such that n mod 4 == 3?",
110 "content": "Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024.",
111111 }
112112 ],
113113 )
from line 137
137137 messages: [
138138 {
139139 role: "user",
140 content: "Are there an infinite number of prime numbers such that n mod 4 == 3?"
140 content: "Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024."
141141 }
142142 ]
143143 });
from line 169
169169 new()
170170 {
171171 Role = Role.User,
172 Content = "Are there an infinite number of prime numbers such that n mod 4 == 3?",
172 Content = "Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024.",
173173 },
174174 ],
175175 });
from line 203
203203 Effort: anthropic.OutputConfigEffortHigh,
204204 },
205205 Messages: []anthropic.MessageParam{
206 anthropic.NewUserMessage(anthropic.NewTextBlock("Are there an infinite number of prime numbers such that n mod 4 == 3?")),
206 anthropic.NewUserMessage(anthropic.NewTextBlock("Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024.")),
207207 },
208208 })
209209 if err != nil {
from line 240
240240 .outputConfig(OutputConfig.builder()
241241 .effort(OutputConfig.Effort.HIGH)
242242 .build())
243 .addUserMessage("Are there an infinite number of prime numbers such that n mod 4 == 3?")
243 .addUserMessage("Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024.")
244244 .build();
245245 
246246 var response = client.messages().create(params);
from line 271
271271 messages: [
272272 [
273273 'role' => 'user',
274 'content' => 'Are there an infinite number of prime numbers such that n mod 4 == 3?',
274 'content' => 'Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024.',
275275 ],
276276 ],
277277 );
from line 297
297297 messages: [
298298 {
299299 role: :user,
300 content: "Are there an infinite number of prime numbers such that n mod 4 == 3?"
300 content: "Find all pairs of positive integers (x, y) such that x^2 - y^2 = 2024."
301301 }
302302 ]
303303 )

models/sonnet-5/overview Changed · +3 / -3 lines

from line 25
2525| Model | Context | Max output | Price / MTok | Latency | Thinking | Default effort | Knowledge cutoff |
2626| :-------------------------------------------------------------------------------- | :------ | :--------- | :----------- | :------- | :------------------- | :------------- | :--------------- |
2727| [Claude Fable 5.1](https://platform.claude.com/docs/en/models/fable-5-1/overview) | 1M | 128K | $10 / $50 | Slower | Adaptive (always on) | `high` | Jun 2026 |
28| [Claude Opus 5](https://platform.claude.com/docs/en/models/opus-5/overview) | 1M | 128K | $5 / $25 | Moderate | Adaptive | `high` | May 2026 |
28| [Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview) | 1M | 128K | $4 / $20 | Moderate | Adaptive (always on) | `medium` | Jun 2026 |
2929| **Claude Sonnet 5** (this model) | 1M | 128K | $2 / $10 | Fast | Adaptive | `high` | Jan 2026 |
3030| [Claude Haiku 4.5](https://platform.claude.com/docs/en/models/haiku-4-5/overview) | 200K | 64K | $1 / $5 | Fastest | Extended | — | Feb 2025 |
3131 
3232* **Context:** 1M tokens is roughly 555k words or 2.5M Unicode characters on the current tokenizer (introduced with Claude Opus 4.7); models before it fit about 750k words in 1M tokens. 200k tokens is roughly 150k words.
33* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
34* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1). See Pricing for the full list.
33* **Max output:** Synchronous Messages API limit. On the Message Batches API, Claude Opus 5.5, Claude Opus 5, Claude Sonnet 5, Claude Opus 4.8, Claude Opus 4.7, Claude Opus 4.6, and Claude Sonnet 4.6 support up to 300k output tokens with the output-300k-2026-03-24 beta header.
34* **Price / MTok:** Input / output, base price per million tokens. Batch API requests are 50% off; prompt caching reads cost 10% of the base input price (2.5% on Claude Fable 5.1 and Claude Mythos 5.1, 5% on Claude Opus 5.5). See Pricing for the full list.
3535* **Latency:** Comparative latency, relative to the current lineup, as published in the models overview. Actual latency depends on prompt length, output length, and thinking effort.
3636* **Thinking:** Adaptive thinking lets the model decide how much to think, steered by effort. Extended thinking is the manual budget\_tokens mode on earlier models.
3737* **Default effort:** The effort parameter’s default on the Claude API. Models without a value don’t support the parameter.

release-notes/overview Changed · +9 / -0 lines

### September 22, 2026

from line 12
1212 For updates to Claude Code, see the [complete CHANGELOG.md](https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md) in the `claude-code` repository.
1313</Tip>
1414 
15### September 22, 2026
16 
17* We've launched **Claude Opus 5.5** (`claude-opus-5-5`), a model for long-running agentic coding and knowledge work. It has a [1M token context window](https://platform.claude.com/docs/en/build-with-claude/context-windows) by default, 128k max output tokens, and always-on [adaptive thinking](https://platform.claude.com/docs/en/build-with-claude/thinking), at $4 / $20 USD per MTok (Claude Opus 5 is $5 / $25). Claude Opus 5.5 is available on the Claude API, [Claude in Amazon Bedrock](https://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrock), [Claude Platform on AWS](https://platform.claude.com/docs/en/build-with-claude/claude-platform-on-aws), [Claude on Google Cloud](https://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-ai), and [Claude in Microsoft Foundry](https://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundry). See [What's new in Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/whats-new-opus-5-5) for capabilities, API changes, and migration guidance.
18* On Claude Opus 5.5, thinking can't be disabled: `thinking: {"type": "disabled"}` and `thinking: {"type": "enabled", ...}` return a 400 error. Omit the `thinking` field and control thinking depth with the [effort parameter](https://platform.claude.com/docs/en/build-with-claude/effort). `tool_choice` types `any` and `tool` also return a 400 error, as on Claude Fable 5.1; use `auto` with [strict tool use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/strict-tool-use). On the Claude API and Google Cloud, [computer use](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool) on this model requires the `computer_toolset_20260801` toolset and the earlier `computer_20251124` tool returns a 400 error; on Amazon Bedrock, `computer_20251124` keeps working. See the [migration guide](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide#migrating-from-claude-opus-5).
19 
20- [Fast mode](https://platform.claude.com/docs/en/build-with-claude/fast-mode) (research preview) is available for Claude Opus 5.5 on the Claude API.
21 
22* Tools can now be defined inside a [mid-conversation system message](https://platform.claude.com/docs/en/build-with-claude/mid-conversation-system-messages#define-tools-in-a-message-beta), in beta on the Claude API with the `inline-tools-2026-09-15` beta header. A `tool_addition` block can carry the tool's full definition (`tool: {"type": "tool_definition", "definition": {...}}`), so you can add a tool, change its schema, or move a server tool to a newer version without editing `tools` or invalidating the prompt cache. The same header covers adding and removing tools by reference. With the MCP connector's `mcp-client-2026-09-15` beta header as well, the definition can be an MCP toolset, and a response records each server's fetched tool list in an `mcp_tool_listing` block, which pins that list when you send it back.
23 
1524### September 18, 2026
1625 
1726* The [Compliance API](https://platform.claude.com/docs/en/manage-claude/compliance-api) local session endpoints now also return transcripts of Claude in Chrome sessions (`product_surface` value `claude_in_chrome`), in beta for Claude Enterprise organizations, with your existing Compliance Access Key and the `read:compliance_user_data` scope. See [Sessions on users' machines](https://platform.claude.com/docs/en/manage-claude/compliance-sessions#retrieve-local-sessions).

test-and-evaluate/develop-tests Changed · +72 / -72 lines

from line 149
149149 
150150 def get_completion(prompt: str):
151151 message = client.messages.create(
152 model="claude-opus-5",
152 model="claude-opus-5-5",
153153 max_tokens=50,
154154 messages=[{"role": "user", "content": prompt}],
155155 )
from line 192
192192 
193193 async function getCompletion(prompt: string): Promise<string> {
194194 const message = await client.messages.create({
195 model: "claude-opus-5",
195 model: "claude-opus-5-5",
196196 max_tokens: 50,
197197 messages: [{ role: "user", content: prompt }]
198198 });
from line 234
234234 {
235235 var message = await client.Messages.Create(new MessageCreateParams
236236 {
237 Model = Model.ClaudeOpus5,
237 Model = Model.ClaudeOpus5_5,
238238 MaxTokens = 50,
239239 Messages = [new() { Role = Role.User, Content = prompt }],
240240 });
from line 304
304304 
305305 func getCompletion(prompt string) string {
306306 message, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
307 Model: anthropic.ModelClaudeOpus5,
307 Model: anthropic.ModelClaudeOpus5_5,
308308 MaxTokens: 50,
309309 Messages: []anthropic.MessageParam{
310310 anthropic.NewUserMessage(anthropic.NewTextBlock(prompt)),
from line 357
357357 
358358 String getCompletion(String prompt) {
359359 var params = MessageCreateParams.builder()
360 .model(Model.CLAUDE_OPUS_5)
360 .model(Model.CLAUDE_OPUS_5_5)
361361 .maxTokens(50L)
362362 .addUserMessage(prompt)
363363 .build();
from line 397
397397 function getCompletion(Client $client, string $prompt): string
398398 {
399399 $message = $client->messages->create(
400 model: Model::CLAUDE_OPUS_5,
400 model: Model::CLAUDE_OPUS_5_5,
401401 maxTokens: 50,
402402 messages: [
403403 [
from line 457
457457 
458458 def get_completion(client, prompt)
459459 message = client.messages.create(
460 model: Anthropic::Model::CLAUDE_OPUS_5,
460 model: Anthropic::Model::CLAUDE_OPUS_5_5,
461461 max_tokens: 50,
462462 messages: [
463463 {
from line 526
526526 
527527 def get_completion(prompt: str):
528528 message = client.messages.create(
529 model="claude-opus-5",
529 model="claude-opus-5-5",
530530 max_tokens=2048,
531531 messages=[{"role": "user", "content": prompt}],
532532 )
from line 581
581581 
582582 async function getCompletion(prompt: string): Promise<string> {
583583 const message = await client.messages.create({
584 model: "claude-opus-5",
584 model: "claude-opus-5-5",
585585 max_tokens: 2048,
586586 messages: [{ role: "user", content: prompt }]
587587 });
from line 668
668668 
669669 def get_completion(prompt: str):
670670 message = client.messages.create(
671 model="claude-opus-5",
671 model="claude-opus-5-5",
672672 max_tokens=1024,
673673 messages=[{"role": "user", "content": prompt}],
674674 )
from line 713
713713 
714714 async function getCompletion(prompt: string): Promise<string> {
715715 const message = await client.messages.create({
716 model: "claude-opus-5",
716 model: "claude-opus-5-5",
717717 max_tokens: 1024,
718718 messages: [{ role: "user", content: prompt }]
719719 });
from line 781
781781 {
782782 var message = await client.Messages.Create(new MessageCreateParams
783783 {
784 Model = Model.ClaudeOpus5,
784 Model = Model.ClaudeOpus5_5,
785785 MaxTokens = 1024,
786786 Messages = [new() { Role = Role.User, Content = prompt }],
787787 });
from line 879
879879 
880880 func getCompletion(prompt string) string {
881881 message, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
882 Model: anthropic.ModelClaudeOpus5,
882 Model: anthropic.ModelClaudeOpus5_5,
883883 MaxTokens: 1024,
884884 Messages: []anthropic.MessageParam{
885885 anthropic.NewUserMessage(anthropic.NewTextBlock(prompt)),
from line 965
965965 
966966 String getCompletion(String prompt) {
967967 var params = MessageCreateParams.builder()
968 .model(Model.CLAUDE_OPUS_5)
968 .model(Model.CLAUDE_OPUS_5_5)
969969 .maxTokens(1024L)
970970 .addUserMessage(prompt)
971971 .build();
from line 1032
10321032 function getCompletion(Client $client, string $prompt): string
10331033 {
10341034 $message = $client->messages->create(
1035 model: Model::CLAUDE_OPUS_5,
1035 model: Model::CLAUDE_OPUS_5_5,
10361036 maxTokens: 1024,
10371037 messages: [
10381038 [
from line 1116
11161116 
11171117 def get_completion(client, prompt)
11181118 message = client.messages.create(
1119 model: Anthropic::Model::CLAUDE_OPUS_5,
1119 model: Anthropic::Model::CLAUDE_OPUS_5_5,
11201120 max_tokens: 1024,
11211121 messages: [
11221122 {
from line 1191
11911191 
11921192 def get_completion(prompt: str):
11931193 message = client.messages.create(
1194 model="claude-opus-5",
1194 model="claude-opus-5-5",
11951195 max_tokens=2048,
11961196 messages=[{"role": "user", "content": prompt}],
11971197 )
from line 1207
12071207 
12081208 # Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
12091209 response = client.messages.create(
1210 model="claude-opus-5",
1210 model="claude-opus-5-5",
12111211 max_tokens=50,
12121212 messages=[{"role": "user", "content": tone_prompt}],
12131213 )
from line 1248
12481248 
12491249 async function getCompletion(prompt: string): Promise<string> {
12501250 const message = await client.messages.create({
1251 model: "claude-opus-5",
1251 model: "claude-opus-5-5",
12521252 max_tokens: 2048,
12531253 messages: [{ role: "user", content: prompt }]
12541254 });
from line 1265
12651265 
12661266 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
12671267 const response = await client.messages.create({
1268 model: "claude-opus-5",
1268 model: "claude-opus-5-5",
12691269 max_tokens: 50,
12701270 messages: [{ role: "user", content: tonePrompt }]
12711271 });
from line 1307
13071307 {
13081308 var message = await client.Messages.Create(new MessageCreateParams
13091309 {
1310 Model = Model.ClaudeOpus5,
1310 Model = Model.ClaudeOpus5_5,
13111311 MaxTokens = 2048,
13121312 Messages = [new() { Role = Role.User, Content = prompt }],
13131313 });
from line 1327
13271327 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
13281328 var response = await client.Messages.Create(new MessageCreateParams
13291329 {
1330 Model = Model.ClaudeOpus5,
1330 Model = Model.ClaudeOpus5_5,
13311331 MaxTokens = 50,
13321332 Messages = [new() { Role = Role.User, Content = tonePrompt }],
13331333 });
from line 1388
13881388 
13891389 func getCompletion(prompt string) string {
13901390 message, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
1391 Model: anthropic.ModelClaudeOpus5,
1391 Model: anthropic.ModelClaudeOpus5_5,
13921392 MaxTokens: 2048,
13931393 Messages: []anthropic.MessageParam{
13941394 anthropic.NewUserMessage(anthropic.NewTextBlock(prompt)),
from line 1409
14091409 
14101410 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
14111411 response, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
1412 Model: anthropic.ModelClaudeOpus5,
1412 Model: anthropic.ModelClaudeOpus5_5,
14131413 MaxTokens: 50,
14141414 Messages: []anthropic.MessageParam{
14151415 anthropic.NewUserMessage(anthropic.NewTextBlock(tonePrompt)),
from line 1461
14611461 
14621462 String getCompletion(String prompt) {
14631463 var params = MessageCreateParams.builder()
1464 .model(Model.CLAUDE_OPUS_5)
1464 .model(Model.CLAUDE_OPUS_5_5)
14651465 .maxTokens(2048L)
14661466 .addUserMessage(prompt)
14671467 .build();
from line 1478
14781478 
14791479 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
14801480 var params = MessageCreateParams.builder()
1481 .model(Model.CLAUDE_OPUS_5)
1481 .model(Model.CLAUDE_OPUS_5_5)
14821482 .maxTokens(50L)
14831483 .addUserMessage(tonePrompt)
14841484 .build();
from line 1512
15121512 function getCompletion(Client $client, string $prompt): string
15131513 {
15141514 $message = $client->messages->create(
1515 model: Model::CLAUDE_OPUS_5,
1515 model: Model::CLAUDE_OPUS_5_5,
15161516 maxTokens: 2048,
15171517 messages: [
15181518 [
from line 1536
15361536 
15371537 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
15381538 $response = $client->messages->create(
1539 model: Model::CLAUDE_OPUS_5,
1539 model: Model::CLAUDE_OPUS_5_5,
15401540 maxTokens: 50,
15411541 messages: [
15421542 [
from line 1590
15901590 
15911591 def get_completion(client, prompt)
15921592 message = client.messages.create(
1593 model: Anthropic::Model::CLAUDE_OPUS_5,
1593 model: Anthropic::Model::CLAUDE_OPUS_5_5,
15941594 max_tokens: 2048,
15951595 messages: [
15961596 {
from line 1613
16131613 
16141614 # Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
16151615 response = client.messages.create(
1616 model: Anthropic::Model::CLAUDE_OPUS_5,
1616 model: Anthropic::Model::CLAUDE_OPUS_5_5,
16171617 max_tokens: 50,
16181618 messages: [
16191619 {
from line 1663
16631663 
16641664 def get_completion(prompt: str):
16651665 message = client.messages.create(
1666 model="claude-opus-5",
1666 model="claude-opus-5-5",
16671667 max_tokens=1024,
16681668 messages=[{"role": "user", "content": prompt}],
16691669 )
from line 1687
16871687 
16881688 # Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
16891689 response = client.messages.create(
1690 model="claude-opus-5",
1690 model="claude-opus-5-5",
16911691 max_tokens=50,
16921692 messages=[{"role": "user", "content": binary_prompt}],
16931693 )
from line 1735
17351735 
17361736 async function getCompletion(prompt: string): Promise<string> {
17371737 const message = await client.messages.create({
1738 model: "claude-opus-5",
1738 model: "claude-opus-5-5",
17391739 max_tokens: 1024,
17401740 messages: [{ role: "user", content: prompt }]
17411741 });
from line 1764
17641764 
17651765 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
17661766 const response = await client.messages.create({
1767 model: "claude-opus-5",
1767 model: "claude-opus-5-5",
17681768 max_tokens: 50,
17691769 messages: [{ role: "user", content: binaryPrompt }]
17701770 });
from line 1803
18031803 {
18041804 var message = await client.Messages.Create(new MessageCreateParams
18051805 {
1806 Model = Model.ClaudeOpus5,
1806 Model = Model.ClaudeOpus5_5,
18071807 MaxTokens = 1024,
18081808 Messages = [new() { Role = Role.User, Content = prompt }],
18091809 });
from line 1833
18331833 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
18341834 var response = await client.Messages.Create(new MessageCreateParams
18351835 {
1836 Model = Model.ClaudeOpus5,
1836 Model = Model.ClaudeOpus5_5,
18371837 MaxTokens = 50,
18381838 Messages = [new() { Role = Role.User, Content = binaryPrompt }],
18391839 });
from line 1899
18991899 
19001900 func getCompletion(prompt string) string {
19011901 message, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
1902 Model: anthropic.ModelClaudeOpus5,
1902 Model: anthropic.ModelClaudeOpus5_5,
19031903 MaxTokens: 1024,
19041904 Messages: []anthropic.MessageParam{
19051905 anthropic.NewUserMessage(anthropic.NewTextBlock(prompt)),
from line 1929
19291929 
19301930 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
19311931 response, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
1932 Model: anthropic.ModelClaudeOpus5,
1932 Model: anthropic.ModelClaudeOpus5_5,
19331933 MaxTokens: 50,
19341934 Messages: []anthropic.MessageParam{
19351935 anthropic.NewUserMessage(anthropic.NewTextBlock(binaryPrompt)),
from line 1980
19801980 
19811981 String getCompletion(String prompt) {
19821982 var params = MessageCreateParams.builder()
1983 .model(Model.CLAUDE_OPUS_5)
1983 .model(Model.CLAUDE_OPUS_5_5)
19841984 .maxTokens(1024L)
19851985 .addUserMessage(prompt)
19861986 .build();
from line 2006
20062006 
20072007 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
20082008 var params = MessageCreateParams.builder()
2009 .model(Model.CLAUDE_OPUS_5)
2009 .model(Model.CLAUDE_OPUS_5_5)
20102010 .maxTokens(50L)
20112011 .addUserMessage(binaryPrompt)
20122012 .build();
from line 2044
20442044 function getCompletion(Client $client, string $prompt): string
20452045 {
20462046 $message = $client->messages->create(
2047 model: Model::CLAUDE_OPUS_5,
2047 model: Model::CLAUDE_OPUS_5_5,
20482048 maxTokens: 1024,
20492049 messages: [
20502050 [
from line 2077
20772077 
20782078 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
20792079 $response = $client->messages->create(
2080 model: Model::CLAUDE_OPUS_5,
2080 model: Model::CLAUDE_OPUS_5_5,
20812081 maxTokens: 50,
20822082 messages: [
20832083 [
from line 2133
21332133 
21342134 def get_completion(client, prompt)
21352135 message = client.messages.create(
2136 model: Anthropic::Model::CLAUDE_OPUS_5,
2136 model: Anthropic::Model::CLAUDE_OPUS_5_5,
21372137 max_tokens: 1024,
21382138 messages: [
21392139 {
from line 2163
21632163 
21642164 # Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
21652165 response = client.messages.create(
2166 model: Anthropic::Model::CLAUDE_OPUS_5,
2166 model: Anthropic::Model::CLAUDE_OPUS_5_5,
21672167 max_tokens: 50,
21682168 messages: [
21692169 {
from line 2242
22422242 
22432243 def get_completion(conversation: list):
22442244 message = client.messages.create(
2245 model="claude-opus-5",
2245 model="claude-opus-5-5",
22462246 max_tokens=1024,
22472247 messages=conversation,
22482248 )
from line 2261
22612261 
22622262 # Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
22632263 response = client.messages.create(
2264 model="claude-opus-5",
2264 model="claude-opus-5-5",
22652265 max_tokens=50,
22662266 messages=[{"role": "user", "content": ordinal_prompt}],
22672267 )
from line 2326
23262326 
23272327 async function getCompletion(conversation: Anthropic.MessageParam[]): Promise<string> {
23282328 const message = await client.messages.create({
2329 model: "claude-opus-5",
2329 model: "claude-opus-5-5",
23302330 max_tokens: 1024,
23312331 messages: conversation
23322332 });
from line 2353
23532353 
23542354 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
23552355 const response = await client.messages.create({
2356 model: "claude-opus-5",
2356 model: "claude-opus-5-5",
23572357 max_tokens: 50,
23582358 messages: [{ role: "user", content: ordinalPrompt }]
23592359 });
from line 2407
24072407 {
24082408 var message = await client.Messages.Create(new MessageCreateParams
24092409 {
2410 Model = Model.ClaudeOpus5,
2410 Model = Model.ClaudeOpus5_5,
24112411 MaxTokens = 1024,
24122412 Messages = [.. conversation.Select(turn => new MessageParam
24132413 {
from line 2436
24362436 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
24372437 var response = await client.Messages.Create(new MessageCreateParams
24382438 {
2439 Model = Model.ClaudeOpus5,
2439 Model = Model.ClaudeOpus5_5,
24402440 MaxTokens = 50,
24412441 Messages = [new() { Role = Role.User, Content = ordinalPrompt }],
24422442 });
from line 2521
25212521 
25222522 func getCompletion(conversation []turn) string {
25232523 message, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
2524 Model: anthropic.ModelClaudeOpus5,
2524 Model: anthropic.ModelClaudeOpus5_5,
25252525 MaxTokens: 1024,
25262526 Messages: toMessageParams(conversation),
25272527 })
from line 2546
25462546 
25472547 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
25482548 response, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
2549 Model: anthropic.ModelClaudeOpus5,
2549 Model: anthropic.ModelClaudeOpus5_5,
25502550 MaxTokens: 50,
25512551 Messages: []anthropic.MessageParam{
25522552 anthropic.NewUserMessage(anthropic.NewTextBlock(ordinalPrompt)),
from line 2607
26072607 
26082608 String getCompletion(List<Turn> conversation) {
26092609 var builder = MessageCreateParams.builder()
2610 .model(Model.CLAUDE_OPUS_5)
2610 .model(Model.CLAUDE_OPUS_5_5)
26112611 .maxTokens(1024L);
26122612 for (var turn : conversation) {
26132613 if (turn.role().equals("user")) {
from line 2635
26352635 
26362636 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
26372637 var params = MessageCreateParams.builder()
2638 .model(Model.CLAUDE_OPUS_5)
2638 .model(Model.CLAUDE_OPUS_5_5)
26392639 .maxTokens(50L)
26402640 .addUserMessage(ordinalPrompt)
26412641 .build();
from line 2681
26812681 function getCompletion(Client $client, array $conversation): string
26822682 {
26832683 $message = $client->messages->create(
2684 model: Model::CLAUDE_OPUS_5,
2684 model: Model::CLAUDE_OPUS_5_5,
26852685 maxTokens: 1024,
26862686 messages: $conversation,
26872687 );
from line 2706
27062706 
27072707 // Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
27082708 $response = $client->messages->create(
2709 model: Model::CLAUDE_OPUS_5,
2709 model: Model::CLAUDE_OPUS_5_5,
27102710 maxTokens: 50,
27112711 messages: [
27122712 [
from line 2772
27722772 
27732773 def get_completion(client, conversation)
27742774 message = client.messages.create(
2775 model: Anthropic::Model::CLAUDE_OPUS_5,
2775 model: Anthropic::Model::CLAUDE_OPUS_5_5,
27762776 max_tokens: 1024,
27772777 messages: conversation
27782778 )
from line 2793
27932793 
27942794 # Generally best practice to use a different model to evaluate than the model used to generate the evaluated output
27952795 response = client.messages.create(
2796 model: Anthropic::Model::CLAUDE_OPUS_5,
2796 model: Anthropic::Model::CLAUDE_OPUS_5_5,
27972797 max_tokens: 50,
27982798 messages: [
27992799 {
from line 2862
28622862 
28632863 def grade_completion(output, golden_answer):
28642864 grader_message = client.messages.create(
2865 model="claude-opus-5",
2865 model="claude-opus-5-5",
28662866 max_tokens=2048,
28672867 messages=[
28682868 {"role": "user", "content": build_grader_prompt(output, golden_answer)}
from line 2894
28942894 
28952895 def get_completion(prompt: str):
28962896 message = client.messages.create(
2897 model="claude-opus-5",
2897 model="claude-opus-5-5",
28982898 max_tokens=1024,
28992899 messages=[{"role": "user", "content": prompt}],
29002900 )
from line 2921
29212921 
29222922 async function gradeCompletion(output: string, goldenAnswer: string): Promise<string> {
29232923 const graderResponse = await client.messages.create({
2924 model: "claude-opus-5",
2924 model: "claude-opus-5-5",
29252925 max_tokens: 2048,
29262926 messages: [{ role: "user", content: buildGraderPrompt(output, goldenAnswer) }]
29272927 });
from line 2946
29462946 
29472947 async function getCompletion(prompt: string): Promise<string> {
29482948 const message = await client.messages.create({
2949 model: "claude-opus-5",
2949 model: "claude-opus-5-5",
29502950 max_tokens: 1024,
29512951 messages: [{ role: "user", content: prompt }]
29522952 });
from line 2980
29802980 {
29812981 var graderResponse = await client.Messages.Create(new MessageCreateParams
29822982 {
2983 Model = Model.ClaudeOpus5,
2983 Model = Model.ClaudeOpus5_5,
29842984 MaxTokens = 2048,
29852985 Messages = [new() { Role = Role.User, Content = BuildGraderPrompt(output, goldenAnswer) }],
29862986 });
from line 3002
30023002 {
30033003 var message = await client.Messages.Create(new MessageCreateParams
30043004 {
3005 Model = Model.ClaudeOpus5,
3005 Model = Model.ClaudeOpus5_5,
30063006 MaxTokens = 1024,
30073007 Messages = [new() { Role = Role.User, Content = prompt }],
30083008 });
from line 3058
30583058 
30593059 func gradeCompletion(output, goldenAnswer string) string {
30603060 graderResponse, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
3061 Model: anthropic.ModelClaudeOpus5,
3061 Model: anthropic.ModelClaudeOpus5_5,
30623062 MaxTokens: 2048,
30633063 Messages: []anthropic.MessageParam{
30643064 anthropic.NewUserMessage(anthropic.NewTextBlock(buildGraderPrompt(output, goldenAnswer))),
from line 3075
30753075 
30763076 func getCompletion(prompt string) string {
30773077 message, err := client.Messages.New(context.Background(), anthropic.MessageNewParams{
3078 Model: anthropic.ModelClaudeOpus5,
3078 Model: anthropic.ModelClaudeOpus5_5,
30793079 MaxTokens: 1024,
30803080 Messages: []anthropic.MessageParam{
30813081 anthropic.NewUserMessage(anthropic.NewTextBlock(prompt)),
from line 3139
31393139 
31403140 String gradeCompletion(String output, String goldenAnswer) {
31413141 var params = MessageCreateParams.builder()
3142 .model(Model.CLAUDE_OPUS_5)
3142 .model(Model.CLAUDE_OPUS_5_5)
31433143 .maxTokens(2048L)
31443144 .addUserMessage(buildGraderPrompt(output, goldenAnswer))
31453145 .build();
from line 3149
31493149 
31503150 String getCompletion(String prompt) {
31513151 var params = MessageCreateParams.builder()
3152 .model(Model.CLAUDE_OPUS_5)
3152 .model(Model.CLAUDE_OPUS_5_5)
31533153 .maxTokens(1024L)
31543154 .addUserMessage(prompt)
31553155 .build();
from line 3184
31843184 function gradeCompletion(Client $client, string $output, string $goldenAnswer): string
31853185 {
31863186 $graderResponse = $client->messages->create(
3187 model: Model::CLAUDE_OPUS_5,
3187 model: Model::CLAUDE_OPUS_5_5,
31883188 maxTokens: 2048,
31893189 messages: [
31903190 [
from line 3213
32133213 function getCompletion(Client $client, string $prompt): string
32143214 {
32153215 $message = $client->messages->create(
3216 model: Model::CLAUDE_OPUS_5,
3216 model: Model::CLAUDE_OPUS_5_5,
32173217 maxTokens: 1024,
32183218 messages: [
32193219 [
from line 3264
32643264 
32653265 def grade_completion(client, output, golden_answer)
32663266 grader_response = client.messages.create(
3267 model: Anthropic::Model::CLAUDE_OPUS_5,
3267 model: Anthropic::Model::CLAUDE_OPUS_5_5,
32683268 max_tokens: 2048,
32693269 messages: [
32703270 {
from line 3290
32903290 
32913291 def get_completion(client, prompt)
32923292 message = client.messages.create(
3293 model: Anthropic::Model::CLAUDE_OPUS_5,
3293 model: Anthropic::Model::CLAUDE_OPUS_5_5,
32943294 max_tokens: 1024,
32953295 messages: [
32963296 {

test-and-evaluate/strengthen-guardrails/handle-streaming-refusals Changed · +8 / -8 lines

from line 66
6666 -H "content-type: application/json" \
6767 -H "x-api-key: $ANTHROPIC_API_KEY" \
6868 -d '{
69 "model": "claude-opus-5",
69 "model": "claude-opus-5-5",
7070 "messages": [{"role": "user", "content": "Hello"}],
7171 "max_tokens": 1024,
7272 "stream": true
from line 95
9595 with client.messages.stream(
9696 max_tokens=1024,
9797 messages=messages + [{"role": "user", "content": "Hello"}],
98 model="claude-opus-5",
98 model="claude-opus-5-5",
9999 ) as stream:
100100 for event in stream:
101101 # Check for refusal in message delta
from line 120
120120 try {
121121 const stream = await client.messages.stream({
122122 messages: [...messages, { role: "user", content: "Hello" }],
123 model: "claude-opus-5",
123 model: "claude-opus-5-5",
124124 max_tokens: 1024
125125 });
126126 
from line 142
142142 
143143 var parameters = new MessageCreateParams
144144 {
145 Model = Model.ClaudeOpus5,
145 Model = Model.ClaudeOpus5_5,
146146 MaxTokens = 1024,
147147 Messages = [new() { Role = Role.User, Content = "Hello" }]
148148 };
from line 184
184184 client := anthropic.NewClient()
185185 
186186 stream := client.Messages.NewStreaming(context.TODO(), anthropic.MessageNewParams{
187 Model: anthropic.ModelClaudeOpus5,
187 Model: anthropic.ModelClaudeOpus5_5,
188188 MaxTokens: 1024,
189189 Messages: []anthropic.MessageParam{
190190 anthropic.NewUserMessage(anthropic.NewTextBlock("Hello")),
from line 220
220220 AnthropicClient client = AnthropicOkHttpClient.fromEnv();
221221 
222222 MessageCreateParams params = MessageCreateParams.builder()
223 .model(Model.CLAUDE_OPUS_5)
223 .model(Model.CLAUDE_OPUS_5_5)
224224 .maxTokens(1024L)
225225 .addUserMessage("Hello")
226226 .build();
from line 261
261261 messages: [
262262 ['role' => 'user', 'content' => 'Hello']
263263 ],
264 model: 'claude-opus-5',
264 model: 'claude-opus-5-5',
265265 );
266266 
267267 foreach ($stream as $event) {
from line 286
286286 
287287 begin
288288 stream = client.messages.stream(
289 model: :"claude-opus-5",
289 model: :"claude-opus-5-5",
290290 max_tokens: 1024,
291291 messages: [{ role: "user", content: "Hello" }]
292292 )

about-claude/model-deprecations Changed · +1 / -0 lines

from line 77
7777| claude-fable-5 | Active | N/A | Not sooner than June 9, 2027 |
7878| claude-mythos-5 | Active | N/A | Not sooner than June 9, 2027 |
7979| claude-mythos-preview | Deprecated | June 9, 2026 | To be announced |
80| claude-opus-5-5 | Active | N/A | Not sooner than September 22, 2027 |
8081| claude-opus-5 | Active | N/A | Not sooner than July 24, 2027 |
8182| claude-opus-4-8 | Active | N/A | Not sooner than May 28, 2027 |
8283| claude-opus-4-7 | Active | N/A | Not sooner than April 16, 2027 |

about-claude/use-case-guides/customer-support-chat Changed · +1 / -1 lines

from line 392
392392```python
393393import time
394394 
395MODEL = "claude-opus-5"
395MODEL = "claude-opus-5-5"
396396 
397397TOOLS = [
398398 {

about-claude/use-case-guides/legal-summarization Changed · +2 / -2 lines

from line 188
188188 
189189 
190190def summarize_document(
191 text, details_to_extract, model="claude-opus-5", max_tokens=1000
191 text, details_to_extract, model="claude-opus-5-5", max_tokens=1000
192192):
193193 # Format the details to extract to be placed within the prompt's context
194194 details_to_extract_str = "\n".join(details_to_extract)
from line 295
295295 
296296 
297297def summarize_long_document(
298 text, details_to_extract, model="claude-opus-5", max_tokens=1000
298 text, details_to_extract, model="claude-opus-5-5", max_tokens=1000
299299):
300300 # Format the details to extract to be placed within the prompt's context
301301 details_to_extract_str = "\n".join(details_to_extract)

agents-and-tools/tool-use/handle-tool-calls Changed · +1 / -1 lines

from line 26
2626 ```json JSON
2727 {
2828 "id": "msg_01Aq9w938a90dw8q",
29 "model": "claude-opus-5",
29 "model": "claude-opus-5-5",
3030 "stop_reason": "tool_use",
3131 "role": "assistant",
3232 "content": [

api/service-tiers Changed · +1 / -1 lines

from line 227
227227 
228228### Supported models
229229 
230Priority Tier is supported on all available Claude models except Claude Fable 5.1, Claude Mythos 5.1, Claude Mythos 5, [Claude Mythos Preview](https://anthropic.com/glasswing), Claude Opus 5, and Claude Sonnet 5.
230Priority Tier is supported on all available Claude models except Claude Fable 5.1, Claude Mythos 5.1, Claude Mythos 5, [Claude Mythos Preview](https://anthropic.com/glasswing), Claude Opus 5.5, Claude Opus 5, and Claude Sonnet 5.
231231 
232232Check the [Models overview](https://platform.claude.com/docs/en/models/overview) for more details on available models.
233233 

build-with-claude/compaction-thinking-blocks Changed · +2 / -1 lines

from line 11
1111 - claude-fable-5
1212 - claude-mythos-5
1313 - claude-mythos-preview
14 - claude-opus-5-5
1415 - claude-opus-5
1516 - claude-opus-4-8
1617 - claude-opus-4-7
from line 56
5556 
5657To add an instruction or change the available tools without touching `system` or `tools`, append the change to `messages`, as described in [Make changes without editing the prefix](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#replace-prefix-edits).
5758 
58Mid-conversation system messages inside the summarized turns are summarized too, so their instructions and tool changes stop applying after the swap. To keep one in force, state it again in a `role: "system"` message directly after the first new `user` turn that follows the kept turns. A system message placed between the block and the kept turns breaks their thinking.
59Mid-conversation system messages inside the summarized turns are summarized too, so their text instructions stop applying after the swap. To keep one in force, state it again in a `role: "system"` message directly after the first new `user` turn that follows the kept turns. Tool changes inside those turns carry over on their own when the compaction request also carries `inline-tools-2026-09-15`: the returned block records their net effect in its `tool_changes` field, so send the block back unmodified. If the block has no `tool_changes` field, restate those tool changes the same way. A system message placed between the block and the kept turns breaks their thinking.
5960 
6061## Check that the kept thinking held
6162