September 29, 2026 · Model Release
Anthropic on Monday released Claude Sonnet 5.5, the second model in its Claude 5.5 family and a mid-tier workhorse aimed at coding, documents, and everyday agent tasks. API pricing is unchanged from Sonnet 5 at $2 per million input tokens and $10 per million output tokens.
The company says Sonnet 5.5 runs more than 30% faster and can cut cost per completed task by up to about 30% through more efficient tool-call batching. Independent tracker Artificial Analysis placed a tested build near 56 on its Intelligence Index, two points behind Opus 5.5 at max effort, with especially strong results on agent and coding work.
The launch continues Anthropic’s rapid 5.5 cadence as it prepares a public listing. Haiku 5.5 is expected next. AWS also listed Sonnet 5.5 on Amazon Bedrock and Claude Platform on AWS the same day, giving enterprise customers an immediate path to test it against existing Claude and competitor endpoints.
For developers, the practical story is not a new price war but higher throughput at the same sticker price. Teams already on Sonnet 5 are being advised to re-check effort settings: the API default is high, while medium is often enough for well-specified coding agents.
Key takeaway — Sonnet 5.5 is an efficiency release: same $2/$10 tokens, faster generation, and lower task cost, positioned as the daily driver while Opus remains the expensive ceiling.
Photo: Unsplash (coding workstation). Sources: Anthropic launch coverage via TechCrunch, AI Weekly, Artificial Analysis, AWS Machine Learning Blog, Sept. 28–29, 2026.
