Support Extended thinking for AWS Bedrock models

Feature Requests Nodes
,
1

The idea is:

It would be great if AWS Bedrock models supported extended thinking mode.

My use case:

Enabling reasoning mode in Claude and Nova models improves their ability to perform complex tasks.

I think it would be beneficial to add this because:

More and more recent models support reasoning, and this capability is essential.

Any resources to support this?

The ChatBedrockConverse function provides support for additional model request configurations through its additional_model_request_fields argument.

from langchain_aws import ChatBedrockConverse

thinking_params= {
    "thinking": {
        "type": "enabled",
        "budget_tokens": 2000
    }
}

model = ChatBedrockConverse(
    model="us.anthropic.claude-3-7-sonnet-20250219-v1:0",
    max_tokens=5000,
    region_name="us-west-2",
    additional_model_request_fields=thinking_params,
)

response = model.invoke("What is the cube root of 50.653?")
print(response.content)

Are you willing to work on this?

TBD