Anthropic launches Claude Sonnet 5.5 with faster output

Anthropic launched Claude Sonnet 5.5, the second model in its Claude 5.5 family, which the company says produces output more than 30% faster and can cost up to 30% less per task for most workloads while per-token pricing stays at $2 per million input tokens and $10 per million output tokens. Anthropic reported Sonnet 5.5 scored 70.6% on Terminal-Bench 4.0 versus 10.3% for Sonnet 5, 55.5% on CursorBench 4.0 versus 34.1%, and 80.1% partial on OSWorld 2.1 versus 57.0%. The model is available on the Claude Platform under the model ID claude-sonnet-5-5 and across AWS, Google Cloud, and Microsoft Azure, with Haiku 5.5 due in the coming weeks.

Anthropic has launched Claude Sonnet 5.5, the second model in its Claude 5.5 family and a direct upgrade to Sonnet 5. The company positions it as its fastest Sonnet yet, built for well-scoped everyday work such as bug fixes, documents, slides, spreadsheets, and design. It is meant to complement Opus 5.5, which remains the stronger choice for complex, open-ended jobs that require sustained judgment, including demanding coding, longer assignments, visual understanding, slide creation, and UI design. Anthropic says Sonnet 5.5 produces output more than 30% faster and can cost up to 30% less per task for most workloads because it requires fewer tokens. Per-token pricing remains unchanged at $2 per million input tokens, $10 per million output tokens, $0.20 per million cache-read tokens, and $2.50 per million cache-write tokens. The reported benchmark gains are substantial. Sonnet 5.5 scored 70.6% on Terminal-Bench 4.0, compared with 10.3% for Sonnet 5, and 55.5% on CursorBench 4.0, compared with 34.1%. At Max effort, it reached 46.2% on FrontierCode 1.1 Main versus 42.4% for its predecessor. Anthropic also says the model is close to Opus 5.5 on GDPval-AA knowledge work and recorded an 80.1% partial score on OSWorld 2.1, up from 57.0% for Sonnet 5. Claude Sonnet 5.5 is available across all platforms, including AWS, Google Cloud, and Microsoft Azure. Developers can select it on the Claude Platform with the model ID claude-sonnet-5-5, and zero data retention is available. Default effort is Medium in Claude Code and Claude https://www.testingcatalog.com/tag/claude/ apps, while Claude Platform defaults to High. Users can adjust effort to trade speed against cost. Those migrating from Sonnet with thinking turned off must switch to the new between tools setting. This is also the first Sonnet release with cyber safeguards and fallback behavior for higher-risk requests, following the approach used for Opus 5.5. Anthropic says biology safeguards remain the same as in Sonnet 5, while new classifiers target reasoning extraction and preserved thinking has been expanded. The company says the narrow controls do not affect routine software development or most life sciences work. The release fills out Anthropic’s Claude 5.5 lineup with a faster, lower-cost option for everyday tasks while keeping Opus at the top for harder work. Haiku 5.5 is due in the coming weeks, extending the family with another tier.