Haven’t tried MCP but in theory, would it be possible to set up a MCP server that has access to reasoning models (i.e., r1) thus adding the agent mode ability to ask reasoning models for implementations steps, bounce ideas around, etc?
Hey, we’ve been looking at adding reasoning models as a tool the agent can call, but this has just been concept work for now, but would have a similar usage!
Keep an eye out on future updates for something like this.
1 Like
This would be awesome. Looking forward to it.