Amp introduced the Dial, which changed the default model from Claude Opus 4.8 to GPT-5.6 Sol and lets users select difficulty levels instead of specific models. The Dial automatically picks the best-performing model for each difficulty tier, tested continuously. Most users adopted the new default without complaint, with 69% of Dial users never changing from the medium setting. Amp found that model choice matters less than task difficulty, context quality, and code review rigor for output quality.