Models

You can choose what model is used during all phases of the Agentic Loop.

Model aliases:

• default — recommended model setting, depending on your account type
• sonnet — latest Sonnet model (currently Sonnet 4.6), for daily coding tasks
• opus — latest Opus model (currently Opus 4.6), for complex reasoning tasks
• haiku — fast and efficient, for simple tasks
• sonnet[1m] — Sonnet with a 1 million token context window, for long sessions
• opusplan — special mode that uses Opus during plan mode, then switches to Sonnet for execution

The three levels of Claude models (intelligence vs. cost tradeoff):

• Claude Opus — really slow, really smart; great for difficult tasks
• Claude Sonnet — all-around well balanced; "the Goldilocks model"
• Claude Haiku — really fast, really dumb; great for general tasks

Related: Agentic Loop, Model Configuration