Batches changedapi/messages/batches
Nearest release: v2.1.296, published 4 hours before this site recorded the change. Shown because the two are within 24 hours of each other. Nothing here says the release caused the edit.
Recorded here
Lines+9added
Lines−29removed
From line
75
where the diff opens
First seen
14 Aug 2026
this site's first read of the page
Recorded edits19to this page, all time
The whole hunk
from line 75, old and new numbered
/
This page is larger than the 256 KiB this site keeps, so one side of the diff below stops where the stored text does.
from line 75
7575
7676 Each input message must be an object with a `role` and `content`. You can specify a single `user`-role message, or you can include multiple `user` and `assistant` messages.
7777
78 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response.
78 If the final message uses the `assistant` role, the response content will continue immediately from the content in that message. This can be used to constrain part of the model's response. This is called prefill. On models that don't support prefill, creating a message that ends with a partial `assistant` response returns a 400 error. See [Prefill not supported](https://platform.claude.com/docs/en/api/errors#prefill-not-supported).
7979
8080 Example with a single `user` message:
8181
from line 93
9393 ]
9494 ```
9595
96 Example with a partially-filled response from Claude:
96 Example with a partially-filled response from Claude, for models that support prefill:
9797
9898 ```json
9999 [
from line 1155
11551155
11561156 - `"claude-haiku-4-5"`
11571157
1158 Fastest model with near-frontier intelligence
1159
11601158 - `"claude-haiku-4-5-20251001"`
11611159
1162 Fastest model with near-frontier intelligence
1163
11641160 - `"claude-opus-4-5"`
11651161
11661162 Powerful intelligence for long-running agents and coding
from line 1333
13371333
13381334 - `thinking: optional ThinkingConfigParam`
13391335
1340 Configuration for enabling Claude's extended thinking.
1336 Configuration for Claude's thinking.
13411337
1342 When enabled, responses include `thinking` content blocks showing Claude's thinking process before the final answer. Requires a minimum budget of 1,024 tokens and counts towards your `max_tokens` limit.
1338 With `{"type": "adaptive"}`, Claude decides when and how much to think. With `{"type": "enabled"}` (manual extended thinking), you set a `budget_tokens` of at least 1,024. Thinking tokens count toward your `max_tokens` limit.
13431339
1344 See [extended thinking](https://platform.claude.com/docs/en/build-with-claude/extended-thinking) for details.
1340 Which `type` values are accepted, and what happens when you omit `thinking`, depend on the model. See [thinking](https://platform.claude.com/docs/en/build-with-claude/thinking#configuring-thinking) for each model's behavior.
13451341
13461342 - `ThinkingConfigEnabled object`
13471343
from line 1355
13591355
13601356 - `display: optional "summarized" or "omitted" or null`
13611357
1362 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
1358 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
13631359
13641360 - `"summarized"`
13651361
from line 1375
13791375
13801376 - `display: optional "summarized" or "omitted" or null`
13811377
1382 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. Defaults to `summarized`.
1378 Controls how thinking content appears in the response. When set to `summarized`, thinking is returned normally. When set to `omitted`, thinking content is redacted but a signature is returned for multi-turn continuity. The default depends on the model; see [Controlling thinking display](https://platform.claude.com/docs/en/build-with-claude/thinking#controlling-thinking-display).
13831379
13841380 - `"summarized"`
13851381
from line 4816
48204816
48214817 - `"claude-haiku-4-5"`
48224818
4823 Fastest model with near-frontier intelligence
4824
48254819 - `"claude-haiku-4-5-20251001"`
48264820
4827 Fastest model with near-frontier intelligence
4828
48294821 - `"claude-opus-4-5"`
48304822
48314823 Powerful intelligence for long-running agents and coding
from line 6329
63376329
63386330 - `"claude-haiku-4-5"`
63396331
6340 Fastest model with near-frontier intelligence
6341
63426332 - `"claude-haiku-4-5-20251001"`
63436333
6344 Fastest model with near-frontier intelligence
6345
63466334 - `"claude-opus-4-5"`
63476335
63486336 Powerful intelligence for long-running agents and coding
from line 7614
76267614
76277615 - `"claude-haiku-4-5"`
76287616
7629 Fastest model with near-frontier intelligence
7630
76317617 - `"claude-haiku-4-5-20251001"`
76327618
7633 Fastest model with near-frontier intelligence
7634
76357619 - `"claude-opus-4-5"`
76367620
76377621 Powerful intelligence for long-running agents and coding
from line 8675
86918675
86928676 pattern: ^srvtoolu_[a-zA-Z0-9_]+$
86938677
8694 - `ToolSearchToolResultBlock object`
8695
8696 - `type: "tool_search_tool_result"`
8697
8698 default: tool_search_tool_result
8699
8700 - `content: ToolSearchToolResultError or ToolSearchToolSearchResultBlock`
8701
8702 - `ToolSearc
8678 - `ToolSearchToolR
No line in this hunk matches that.