You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on May 15, 2026. It is now read-only.
Repository navigation
This repository was archived by the owner on May 15, 2026. It is now read-only.
OpenRouter reasoning details aren't preserved between conversation turns, breaking reasoning continuity for models like MiniMax-M2
Context (who is affected and when)
Users of Roo Code who use OpenRouter as their API provider with reasoning-capable models are affected during multi-turn conversations. This impacts the quality of responses in complex tasks that require the model to build upon previous reasoning, such as debugging sessions, architectural discussions, or iterative problem-solving workflows.
Desired behavior (conceptual, not technical)
When using reasoning-capable models through OpenRouter, the system should preserve the model's reasoning traces throughout the conversation. This allows the model to reference and build upon its previous reasoning in subsequent responses, maintaining a coherent thought process across multiple turns and improving the quality of assistance for complex, multi-step tasks.
Constraints / preferences (optional)
No response
Request checklist
I've searched existing Issues and Discussions for duplicates
This describes a specific problem with clear context and impact
Activity
feat: preserve reasoning details for OpenRouter models between conver…
piiq commentedon Dec 19, 2025
additional context:
This also impacts models like Gemini 3 Flash