Checked other resources

• This is a feature request, not a bug report or usage question.
• I added a clear and descriptive title that summarizes the feature request.
• I used the GitHub search to find a similar feature request and didn't find it.
• I checked the LangChain documentation and API reference to see if this feature already exists.
• This is not related to the langchain-community package.

Package (Required)

• langchain
• langchain-openai
• langchain-anthropic
• langchain-classic
• langchain-core
• langchain-cli
• langchain-model-profiles
• langchain-tests
• langchain-text-splitters
• langchain-chroma
• langchain-deepseek
• langchain-exa
• langchain-fireworks
• langchain-groq
• langchain-huggingface
• langchain-mistralai
• langchain-nomic
• langchain-ollama
• langchain-perplexity
• langchain-prompty
• langchain-qdrant
• langchain-xai
• Other / not sure / general

Feature Description

Add support for reasoning outputs from OpenAI-compatible models that don't support the Responses API.

Current behavior:

• OpenAI-compatible servers that support reasoning but not Responses API have their reasoning content ignored
• Reasoning content in choices[].message.reasoning or choices[].message.reasoning_content is not extracted

Expected behavior:

• Extract reasoning content from standard Chat Completions API responses
• Include reasoning in AIMessage.content or AIMessageChunk.content

Use Case

Some OpenAI-compatible inference servers (vLLM, custom deployments) support reasoning models but does not support Responses API.

Problem:

• Enterprise self-hosted inference servers with reasoning models
  • For Example:
    • OpenAI Compatible API provided by vLLM
    • OpenAI Compatible API provided by Cohere (Cannot use langchain-cohere because of bug in it, I already reported)
    • Other providers that provide OpenAI Compatible API but does not support Responses API
• Reasoning content is lost and inaccessible through LangChain
  • Must manually parse raw responses to extract reasoning
  • Breaks compatibility with agent frameworks

Example:

llm = ChatOpenAI(
    base_url="https://custom-server.com/v1",
    model="reasoning-model",
    reasoning_effort="medium", 
    use_responses_api=False  # Server doesn't support it
)

response = llm.invoke("Solve this problem...")
# Currently: reasoning is ignored
# Needed: reasoning accessible in response.content or response.content_blocks

Why needed:

• Broader compatibility with OpenAI-compatible servers that implement reasoning
• Many servers support Chat Completions API with reasoning fields but not the full Responses API
• Enable reasoning capabilities across diverse inference server implementations

Proposed Solution

Expectation:

1. Detect reasoning fields in API response:
   • Check choice.message.reasoning or choice.message.reasoning_content
   • Only when use_responses_api=False and reasoning_effort (or reasoning) is provided
2. Include in message content:
   • Add as content block with type="reasoning"
   • Similar format to Responses API handling
3. Maintain backward compatibility:
   • Only process when fields exist in response
   • No change for models without reasoning support

Alternatives Considered

No response

Additional Context

This could be similar to #34011 and #34014