Introducing Claude Sonnet 5.5

Anthropic has announced Claude Sonnet 5.5, the second model in the Claude 5.5 family and a clear upgrade over Claude Sonnet 5. According to the company, Sonnet 5.5 generates outputs over 30% faster and costs up to 30% less for most work compared to its predecessor. While priced identically to Sonnet 5 at $2 per million input tokens and $10 per million output tokens, the new model typically requires fewer tokens and features halved cache read costs at $0.10 per million tokens. Sonnet 5.5 is optimized for well-scoped everyday tasks, fixing bugs, and creating polished documents, slides, and spreadsheets, while complementing the complex reasoning strengths of Claude Opus 5.5. Performance metrics show a dramatic improvement, including a 70.6% score on the Terminal-Bench 4.0 agentic coding evaluation compared to Sonnet 5's 10.3%, and strong results on GDPval-AA. In coding tasks and long-horizon work, early testers found Sonnet 5.5 to be a more efficient and natural collaborative partner, exhibiting improved tool-calling batching. Due to its advanced cybersecurity capabilities, Sonnet 5.5 is the first Sonnet model to launch with specialized cyber safeguards and fallbacks similar to those developed for Anthropic's most capable models, while routine software development remains unaffected. A forthcoming Claude Haiku 5.5 model will complete the family in the coming weeks.