A fast model is still the wrong choice if it cannot fit your context. Make context window and capabilities pass/fail checks; rank latency, output style, and provider availability afterward. #AIModels #LLMSelection https://temprhq.io/models