Sonnet 5.5 targets everyday coding with fewer tokens

Anthropic’s Sonnet 5.5 keeps Sonnet 5’s token prices while aiming to finish tasks with fewer tokens and faster output. It targets bounded coding work and document creation, with effort settings that trade speed against longer reasoning. Anthropic still positions Opus 5.5 above it for complex, open-ended work, so benchmark proximity should not imply interchangeable behaviour.

Anthropic · Sep 28, 2026

Why it matters

Teams can test whether routine agent tasks need a more expensive model.

More context

Try a representative bug-fix evaluation using claude-sonnet-5-5. Standard API pricing is $2 per million input tokens and $10 per million output tokens. Integrations with thinking disabled need the documented between_tools migration.

Keep in mind

The model is released across supported platforms. Efficiency and performance claims come from Anthropic’s evaluations; 1min has not reproduced them. Higher-risk cybersecurity requests can fall back to Sonnet 5.

Original read