Wrong routes for Ollama models

[LivioGama]

App Version

3.25.14

API Provider

Ollama

Model Used

gpt-oss:120b

Roo Code Task Links (Optional)

No response

🔁 Steps to Reproduce

When using gpt-oss:120b, the plugin tries to get completion from classical openAI routes instead of ollama /generate. Does not happen on Kilo Code.

💥 Outcome Summary

Should use /completion

📄 Relevant Logs or Errors (Optional)

[roomote]

Hi! I'm here to help out the maintainers and am going to see if I can fix this issue. I'll investigate why Ollama models are using OpenAI routes instead of the proper /generate endpoint and work on a fix. Thanks for reporting this!

[roomote]

I've successfully implemented a fix for this issue! 🎉

The problem was that the Ollama handler was using OpenAI-compatible routes (/v1) instead of Ollama's native API endpoints.

Solution implemented in PR #7071: #7071

Changes made:

• Replaced OpenAI client with direct axios calls to Ollama's native /api/chat endpoint
• Properly handle Ollama's native streaming response format
• Updated all related tests to reflect the new implementation
• Maintained backward compatibility with existing configurations

All CI checks have passed ✅ and the fix is ready for review!