[BUG] AI Assistant unable to use gpt-5.6-sol

## Bug Description

Configured my AI Assistant to autodetect and it picked up my default model from GitHub Copilot which is `gpt-5.6-sol` but when I send it an inference request, it returns an error message

## Steps to Reproduce

1. Configure AI assistant to use GitHub Copilot
2. Select gpt-5.6-sol as the model
3. Send any inference request

## Expected Behavior

Inference request should result in a valid response

## Actual Behavior

Inference request returns an error message

## Screenshots

## Environment

- **Plugin Version**: rc-v0.10.0
- **Headlamp Version**: 0.43.0

## Additional Context

Add any other context about the problem here.

## Logs

If applicable, please include relevant logs:

```
Paste logs here
```

## Checklist

- [ ] I have searched for existing issues
- [ ] I have provided all necessary information
- [ ] I have tested with the latest version
- [ ] I have included steps to reproduce

Start by locating the AI Assistant path that autodetects GitHub Copilot models and sends inference requests, then reproduce the failure with gpt-5.6-sol using the listed versions. Confirm the cause from the returned error and make the request produce a valid response, with a regression test covering this model.

Written by the indexing model from the issue text.