Claude Opus 5.5 is now available on AWS

AWS says Claude Opus 5.5 is now available on Amazon Bedrock and Claude Platform on AWS, and it’s the first model in the Claude 5.5 family. Anthropic calls it its most capable Opus model, aimed at agentic coding, knowledge work, and long-running tasks. That puts it squarely in the class of models people will lean on for work that doesn’t fit neatly into a single prompt-and-reply exchange.

The headline here is efficiency. Anthropic says Opus 5.5 does more with fewer tokens than Claude Opus 5, and AWS says the pricing changes pass those gains on to customers. Lower per-token prices and much cheaper cache reads are part of the pitch, with the result described as a lower average cost per task than Opus 5. For teams trying to run larger agent workflows without watching the meter like a hawk, that matters.

The model is also designed to talk more clearly while it works. AWS says it surfaces what it did, what it found, and what it needs, which should make long jobs easier to follow. Adaptive thinking is always on, so Opus 5.5 decides how much reasoning a task needs instead of forcing users to set manual thinking budgets. That’s a quiet but important shift: less fiddling, more asking.

There’s a safety angle too. AWS says this is the first Opus model with safety classifiers similar to Claude Fable 5.1 in biology, cyber security, and AI development. Requests will be refused more often than on earlier Opus versions. In other words, this is not just a bigger model; it’s a more guarded one.

AWS says the model is available today through US, EU, AU, JP, and Global CRIS inference profiles on bedrock-runtime, and also in US East (N. Virginia) and AP SouthEast (Melbourne) on bedrock-mantle. It’s available through Claude Platform on AWS in North America as well. For developers, AWS points to the Bedrock console, the AWS CLI, the AWS SDKs, and the Anthropic Messages API as ways to get started.