Feature Request: Real-time Model Toggle
I often start sessions with thinking models (e.g., claude-4-sonnet-thinking) for complex reasoning, then switch to non-thinking models for execution-focused tasks. Currently, I have to stop conversations, change models, and restart with “We just got cut off, please continue.”
Requested Feature:
A toggle switch to seamlessly switch between thinking/non-thinking model variants mid-conversation without interrupting the session.
Use Case:
This would enable more efficient workflows where users can dynamically adjust the model’s reasoning depth based on the current task requirements.