Skip to content

The agent engineering platform.

chore(anthropic): update docstring following sonnet-5 release (#38577)

ccurme8a2f1a9445ed
diff --git a/libs/partners/anthropic/langchain_anthropic/chat_models.py b/libs/partners/anthropic/langchain_anthropic/chat_models.py
index 9def0e668e..edffea935a 100644
--- a/libs/partners/anthropic/langchain_anthropic/chat_models.py
+++ b/libs/partners/anthropic/langchain_anthropic/chat_models.py
@@ -1005,15 +1005,17 @@ class ChatAnthropic(BaseChatModel):
Examples:
- `#!python {"type": "enabled", "budget_tokens": 10_000}` (pre-4.7 models)
- - `#!python {"type": "adaptive"}` (Opus 4.6+)
- - `#!python {"type": "adaptive", "display": "summarized"}` (Opus 4.7+)
+ - `#!python {"type": "adaptive"}` (Opus 4.6+, Sonnet 5)
+ - `#!python {"type": "adaptive", "display": "summarized"}` (Opus 4.7+, Sonnet 5)
+ - `#!python {"type": "disabled"}` (Sonnet 5, where adaptive thinking is
+ on by default)
- !!! note "Claude Opus 4.7"
+ !!! note "Claude Opus 4.7+ and Sonnet 5"
- `budget_tokens` is removed on Opus 4.7 — use `{"type": "adaptive"}`
- with `output_config.effort` to control reasoning effort. Set `display`
- to `"summarized"` to receive summarized reasoning in the response
- (default is `"omitted"`).
+ `budget_tokens` is removed on these models — use `{"type": "adaptive"}`
+ with `output_config.effort` to control reasoning effort. The default
+ `display` is `"omitted"`; set it to `"summarized"` to receive
+ summarized reasoning in the response.
"""
output_config: dict[str, Any] | None = None