Start with the recommended fast model. Everything runs on your device, and you can switch anytime.
Instant responses. Perfect for everyday conversations.
Better reasoning with a small tradeoff in speed.
Experimental models for enthusiasts. May be slower on some devices.