Support for Sonnet-3.7 thinking process through init_chat_model

[smartinezbragado]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

I would love to be able to use the thinking process of sonnet 3.7 through the init_chat_model feature - It's a game changer for our agents

Motivation

I would love to be able to use the thinking process of sonnet 3.7 through the init_chat_model feature - It's a game changer for our agents

Proposal (If applicable)

No response

[ccurme]

This is already supported:

from langchain.chat_models import init_chat_model

llm = init_chat_model(
    "anthropic:claude-3-7-sonnet-latest",
    max_tokens=5_000,
    thinking={"type": "enabled", "budget_tokens": 2_000},
)

response = llm.invoke("What is 3*3?")
response.content

See docs for ChatAnthropic.

Let me know if this isn't what you meant!