[BUG] Openai - The language model did not provide any assistant messages

[politsin]

Problem (one or two sentences)

Date/time: 2025-12-25T08:53:48.612Z
Extension version: 3.37.1
Provider: openai-native
Model: gpt-5.2

Unexpected API Response: The language model did not provide any assistant messages. This may indicate an issue with the API or the model's output.

Context (who is affected and when)

ChatGPT-openai

Reproduction steps

• select ChatGPT
• send message to chat

Expected result

message

Actual result

error

Variations tried (optional)

No response

App Version

3.37.1

API Provider (optional)

OpenAI

Model Used (optional)

gpt-5.2

Roo Code Task Links (optional)

No response

Relevant logs or errors (optional)

[roomote]

I'd like to help with this issue. After investigating the codebase, I have a few clarifying questions:

1. Model availability: Are you using a custom OpenAI-compatible endpoint, or the official OpenAI API? The model "gpt-5.2" is defined in the Roo Code model configuration, but I want to confirm this model exists in your OpenAI account.

2. Error details: The error "The language model did not provide any assistant messages" typically occurs when the API returns a response without text or tool call content. This could happen if:

   • The model name is not recognized by the API (resulting in an error that gets handled as an empty response)
   • The API returns an error that is being silently caught
   • The model genuinely returns no content

3. Debug logs: Could you enable verbose logging and share any console output or error messages? In VS Code, you can open the Developer Tools (Help > Toggle Developer Tools) and check the Console tab for any errors.

4. API key: Is your OpenAI API key valid and does it have access to the model you are trying to use?

Please tag @roomote in your response with additional details, and I'll be happy to help further diagnose or fix this issue.

[politsin]

I've figured out the problem - when using the checkbox "Use a custom base URL," the host format changed.

Previously:

https://my-host.example

Now:

https://amy-host.example/v1

[politsin]

I would suggest explicitly stating the base URL format when the checkbox is used. And provide a warning when changing it. Preferably, of course, allow both formats to be used: both https://my-host.example/ and https://my-host.example/v1.