Continue CLI does not work with external OpenAI-compatible model (Error: 400 status code (no body))

Before submitting your bug report

• I've tried finding an answer on the Continue docs site
• I'm not able to find an open issue that reports the same bug
• I've seen the troubleshooting guide on the Continue Docs

Relevant environment info

- OS: LinuxMint DE
- Continue version: 1.5.47
- IDE version: --
- Model: External, OpenAI compatible
- config:
  
  - name: MyOpanAICompatibleModel
    provider: openai
    model: AUTODETECT
    apiBase: https://myurl/.../api
    apiKey: ${{ secrets.MY_KEY }}
    roles:
      - chat
      - edit
      - apply

Description

When I configure cn to use an OpenAI-compatible model run on my server (which works perfectly in VSCod(ium) continue plugin), I get: Error: 400 status code (no body)

There are lots of such reports in the issues, but they all have been close for lack of activity...
https://github.com/continuedev/continue/issues?q=is%3Aissue%20state%3Aclosed%20Error%3A%20400%20status%20code%20(no%20body)

This makes the tool completely unusable... please can someone look at it?

To reproduce

1. install cn CLI
2. Configure it to use any OpenAI-compatible model (e.g. through Ollama)
3. Experience the 400 error

Log output

Start by reproducing the issue through the Continue CLI with the provided OpenAI-compatible configuration and an Ollama-compatible server, then trace where the 400 status code is surfaced. Done means the CLI works with a supported external OpenAI-compatible model or reports a useful, actionable error instead of an empty response.