"Encrypted content is not supported with this model" for gpt-5-chat-latest

[andrzej-pomirski-yohana]

Initial Checks

• I confirm that I'm using the latest version of Pydantic AI
• I confirm that I searched for my issue in https://github.com/pydantic/pydantic-ai/issues before opening this issue

Description

I am trying to use the Responses API to run my model to potentially gain some performance (need to squeeze everything I can from the models).

The code (in dependency-injector for Python) of my container is:

class ExampleContainer(DeclarativeContainer):
    llm_provider = Singleton(LiteLLMProvider, api_base=llm_base, api_key=llm_key)
    model = Singleton(
        OpenAIResponsesModel,
        model_name="openai/gpt-5-chat-latest",
        provider=llm_provider
    )

This results in

pydantic_ai.exceptions.ModelHTTPError: status_code: 400, model_name: openai/gpt-5-chat-latest, body: {'message': 'litellm.BadRequestError: OpenAIException - {\n  "error": {\n    "message": "Encrypted content is not supported with this model.",\n    "type": "invalid_request_error",\n    "param": "include",\n    "code": null\n  }\n}No fallback model group found for original model_group=openai/gpt-5-chat-latest. Fallbacks=[]. Received Model Group=openai/gpt-5-chat-latest\nAvailable Model Group Fallbacks=None\nError doing the fallback: litellm.BadRequestError: OpenAIException - {\n  "error": {\n    "message": "Encrypted content is not supported with this model.",\n    "type": "invalid_request_error",\n    "param": "include",\n    "code": null\n  }\n}No fallback model group found for original model_group=openai/gpt-5-chat-latest. Fallbacks=[] LiteLLM Retried: 1 times, LiteLLM Max Retries: 2', 'type': None, 'param': None, 'code': '400'}

However, if I add

        profile=replace(
            OpenAIModelProfile.from_profile(openai_model_profile("gpt-5-chat-latest")),
            openai_supports_encrypted_reasoning_content=False,
        ),

to OpenAIResponsesModel, my model works.

The same code (without custom profile) also works for gpt-5, so I assume that the issue is in the gpt-5-chat-latest model not working with the field.

Example Code

Python, Pydantic AI & LLM client version

Using pydantic-ai-slim 1.9.1 with Python 3.12. OpenAI lib version 2.7.0. LiteLLM server 1.78.5.

[RaymanP02]

Found that pydantics profile was assuming that all gpt-5 models are reasoning hence it erroring on the non-reasoning model.