Anthropic's Claude Opus 5.5, the first release in the Claude 5.5 family, is now available on Amazon Bedrock and Claude Platform on AWS, according to the AWS Machine Learning Blog.
The model is positioned for agentic coding, knowledge work, and long-running tasks. Anthropic says it does more with fewer tokens than Opus 5, and AWS has paired that with lower per-token pricing and cheaper cache reads, producing an average lower cost per task. Opus 5.5 also surfaces what it did, what it found, and what it needs as it works, making long agent sessions easier to follow. Adaptive thinking is always on, replacing manual thinking-budget controls with a single "effort" setting.
Opus 5.5 is the first Opus model to ship with safety classifiers — similar to those on Claude Fable 5.1 — covering biology, cyber security, and AI development, and AWS notes it will refuse requests more often than earlier Opus versions.
In practice, Anthropic points to two use cases: software development, where longer sessions benefit from clearer, more explainable communication, and knowledge work, where the model needs fewer corrections while producing or editing long documents and reports.
The model is accessible via US, EU, AU, JP, and Global Geo CRIS inference profiles on bedrock-runtime, and runs directly in US East (N. Virginia) and AP Southeast (Melbourne) on bedrock-mantle. It's also available through Claude Platform on AWS in North America. Developers can call it through the Bedrock Invoke and Converse APIs, or through the Anthropic Messages API via the Anthropic SDK.