Claude Sonnet 5.5

What changed

Anthropic announced Claude Sonnet 5.5 on September 28. It positions the model for well-scoped coding tasks and professional documents, with Opus 5.5 remaining its stronger choice for complex, open-ended work requiring sustained judgment.

Price versus task cost

Base API rates are $2 per million input tokens and $10 per million output tokens, with cache reads at $0.20 per million tokens. Those are the same base prices as Sonnet 5. Anthropic’s claim of up to 30% lower cost per task comes from using fewer tokens in its tests, not a reduction in the listed input and output prices.

The company also reports output generation more than 30% faster than Sonnet 5. These speed and task-cost claims are provider measurements and should be checked on the workload you actually run.

Migration details

The announced API ID is claude-sonnet-5-5. Anthropic says the model is available on its platform and through Amazon Web Services, Google Cloud and Microsoft Azure. Deployments using thinking off need to inspect the documented between_tools setting before migrating.

The announcement says higher-risk cybersecurity tasks can fall back to Sonnet 5 under its safeguards. That behavior matters for applications expecting a fixed model throughout a workflow.

Source

• Anthropic announcement, September 28