Model level rate limit settings

[sati-bodhi]

Is it possible to configure a profile's rate limit on a model by model basis? This would make it even more convenient to fine-tune how different models fit into the workflow for highest efficiency. The reason is because model providers usually have different rate limits for different models. Gemini 2.5 Pro has higher restrictions than 2.5 flash, for example. Without this granularity, we would have to either work on the lowest bar or tune the rate limit every time we switch models.