Anthropic ships Claude Sonnet 5.5 at Sonnet 5 prices
The second model in the 5.5 family: more than 30% faster output, up to 30% less per task, the same $2 / $10 per million tokens, and 70.6% on Terminal-Bench 4.0 against Sonnet 5's 10.3%. Two points below Opus 5.5 on GDPval-AA.
Claude Sonnet 5.5 (model id claude-sonnet-5-5) arrived on 28 September, six days after Opus 5.5, at the same price as Sonnet 5: $2 per million input tokens and $10 per million output, with cache reads at 20 cents. Anthropic puts output speed over 30% faster than Sonnet 5 and the cost of a typical task up to 30% lower because it uses fewer tokens; on Terminal-Bench 4.0 it scores 70.6% where Sonnet 5 scored 10.3%, and on GDPval-AA it lands two points below Opus 5.5 and 395 above Sonnet 5. Anthropic pitches it for well-scoped everyday work, bug fixing, and documents, slides and spreadsheets, with Opus 5.5 kept for open-ended work that needs sustained judgment. It is the first Sonnet with the cybersecurity safeguards of the Opus tier, carries preserved thinking and the classifiers against reasoning extraction, and hands higher-risk security requests back to Sonnet 5. Available on the Claude Platform, AWS, Google Cloud and Microsoft Azure; the default effort is medium in Claude Code and high on the Platform. Haiku 5.5 is announced for the coming weeks.
Read the source