AWS announced in its Machine Learning Blog that Claude Sonnet 5.5 is now available on Amazon Bedrock and Claude Platform on AWS. The model targets focused coding and knowledge work, with a lower cost per task than Sonnet 5 for most workloads and faster response speed.
AWS says the model checks its own output against stated requirements when given a feature or bug-fix task, and produces more polished documents and visuals — one-pagers, architecture diagrams, summary slides — than its predecessor. The combination of accuracy on well-scoped work and lower cost is aimed at continuous or at-scale use cases: first response to alerts, always-on agent monitoring, SQL generation, UI/UX testing, and IDE coding agents running under a fixed spend cap. AWS also points to scoped analysis, spreadsheet edits, and routine document tasks across large user populations as fits.
Sonnet 5.5 is positioned to pair with Opus 5.5, released earlier, which AWS recommends for judgment-heavy work: release debugging, multi-PR feature stacks, security review of large pull requests, code migrations, long analyses ending in a finished deliverable, financial research, and contract redlining. Sonnet 5.5 is meant to take over once the approach is already decided and execution is what's left.
On Bedrock, the model runs inside existing AWS controls — IAM for access, CloudTrail for audit, CloudWatch for monitoring, and Bedrock Guardrails — with usage billed through the standard AWS account and data kept within AWS infrastructure via Regional data residency. It's accessible today through the Global CRIS (global.) inference profile on bedrock-runtime, via the Bedrock console Playground, the Anthropic Messages API, or the Invoke/Converse APIs through the AWS CLI and SDK. It is also available through Claude Platform on AWS in North America. AWS provides a getting-started notebook and a sample Python script using Boto3 for teams building against it directly.