Overview
Claude Sonnet 5.5 sits between Opus and Haiku in Anthropic’s September 2026 lineup. Thinking is adaptive and on by default, and the lowest setting, between_tools, turns off up-front thinking for faster responses.
Highlights
- 1M-token context window
- Up to 128K output tokens
- Adaptive thinking with high effort by default
- Text and image input
- Cache reads at $0.10 per 1M tokens
Typical uses
Production assistants, coding tools, agents and high-volume reasoning tasks.
Things to consider
- Setting temperature, top_p or top_k returns an error.
- Several breaking changes from Claude Sonnet 5; see the migration guide.
Specifications are taken from the provider’s official announcement, documentation or model card. Benchmark results depend on evaluation settings; see how AI benchmarks work.