Feature request for product/service
Chat
Describe the request
Currently, the model dropdown (in Chat, Composer, and Ctrl+K) only allows selecting a single model or “default.” When working on different types of tasks throughout the day, I find myself constantly switching between models manually – quick tasks get a fast model, complex architecture decisions get an expert-tier model, etc.
There’s no way to define reusable groups of models or balance usage across a set of models within a tier.
Allow users to define model groups (or “tiers”) in Cursor Settings, and expose them as selectable options alongside individual models in the dropdown.
Example tiers:
Tier Models
Standard Composer 1, Kimi K2, Claude Haiku 4.5, GPT-5.2 Codex, GPT-5.3 Codex, Grok Code Fast
Advanced Claude Haiku 4.5 (Thinking), GPT-5.2, Composer 1.5, Gemini 3 Flash, Gemini 3 Pro
Expert Claude Opus 4.6, Claude Opus 4.6 (Thinking), Claude Sonnet 4.5, GPT-5.2 Codex Extra High
Behavior when a group is selected:
Round-robin or random balancing across the models in the group (one model per request)
If a model in the group is rate-limited or unavailable, automatically fall back to the next one in the group
Optionally allow setting a primary + fallback order within the group
Why This Is Useful
Cost management – Balance between expensive and cheap models within a budget tier
Rate limit resilience – If one model is throttled, the group auto-rotates to the next
Workflow-based selection – Pick “Standard” for boilerplate, “Expert” for complex logic, without remembering individual model names
Team standardization – Teams could share group definitions via workspace settings so everyone uses the same model policies