Support openai image_generation via responses API

[felixhuttmann]

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

Currently (langchain-openai 0.3.17), the responses API as wrapped by ChatOpenAi does not support image generation via the new responses API, i.e. the following is not possible via langchain. Instead it errors on the branch in [1].

ChatOpenAI(model="gpt-4.1-mini").bind_tools([{"type": "image_generation"}]) # error: ValueError: Unsupported function

Example without langchain taken from [2]:

response = client.responses.create(
    model="gpt-4.1-mini",
    input="Generate an image of gray tabby cat hugging an otter with an orange scarf",
    tools=[{"type": "image_generation"}],
)

[1] https://github.com/langchain-ai/langchain/blob/langchain-openai%3D%3D0.3.17/libs/core/langchain_core/utils/function_calling.py#L559
[2] https://platform.openai.com/docs/guides/image-generation?image-generation-model=gpt-image-1#generate-images

Motivation

The new openai response API with the gpt-4.1 model permits multi-turn, conversational editing of images. Langchain should support this because the future of LLMs is multi-modal.

Proposal (If applicable)

No response