Anthropic released Claude Sonnet 5.5 on September 28, positioning it as a faster model for well-scoped coding and office work. The company says it generates output more than 30% faster than Sonnet 5 and costs up to 30% less per task in its testing.

That is an efficiency claim, not a reduction in token prices. Anthropic lists Sonnet 5.5 at $2 per million input tokens and $10 per million output tokens, with cache reads at $0.20 per million. The reported savings come from needing fewer tokens to complete work; actual costs will depend on the task and reasoning settings.

Deployment and migration

Developers can access the model on the Claude Platform as claude-sonnet-5-5. Anthropic also lists availability through Amazon Web Services, Google Cloud and Microsoft Azure. Vercel separately confirmed AI Gateway support under anthropic/claude-sonnet-5.5, including zero data retention.

There is a migration detail for applications that disable thinking: Anthropic says they must switch to the new between_tools setting, which keeps up-front thinking off, before moving to Sonnet 5.5. Claude Code and the Claude apps default to Medium effort, while the Claude Platform defaults to High.

New cybersecurity safeguards

Anthropic says Sonnet 5.5 has cybersecurity capabilities comparable to Opus 5, making it the first Sonnet release with similar cyber safeguards. Higher-risk cybersecurity tasks will visibly fall back to Sonnet 5, while routine bug-finding and fixes remain available.

The launch offers developers another cost-performance option for everyday agent workloads. Anthropic still describes Opus 5.5 as stronger for complex, open-ended work requiring sustained judgment; the published speed and savings figures remain vendor-reported results.