Introducing Claude Sonnet 5.5 on AWS
Amazon Web Services Dani Mitchell ● Covered by 12 sources
AWS added Claude Sonnet 5.5 to Bedrock and Claude Platform on AWS. It’s built for faster, cheaper focused work while keeping data inside AWS.
Based on reporting by Amazon Web Services, Dani Mitchell — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
AWS is now offering Claude Sonnet 5.5 on Amazon Bedrock and on Claude Platform on AWS. The pitch is pretty clear: this is the Sonnet model for focused coding and knowledge work, with lower cost per task for most jobs and faster speed than before.
The model seems aimed at work that already has a shape. AWS says it does well on scoped tasks like feature work and bug fixes, and it checks its own output against the requirements it was given. It also does better than Sonnet 5 at polished documents and visuals, including one-pagers, architecture diagrams, and summary slides. That makes it useful when the same kind of job keeps showing up over and over, from alert response and monitoring to SQL generation, UI and UX testing, and coding agents with a fixed spend cap.
AWS is also leaning hard on the enterprise angle. Bedrock keeps data inside AWS infrastructure with Regional data residency, and it plugs into IAM, CloudTrail, CloudWatch, and Bedrock Guardrails. Usage shows up on the AWS bill, which is exactly the sort of sentence that makes procurement people relax and engineers sigh.
The company is pairing Sonnet 5.5 with Opus 5.5 rather than treating them as rivals. Opus 5.5 is for the judgment calls: release debugging, multi-PR feature stacks, security review of large pull requests, code migrations with clear targets, and longer analyses that end in a spreadsheet, deck, or report. Sonnet 5.5 is for the parts where the path is already known and the job is to execute quickly.
Getting started is straightforward enough. You can try it in the Amazon Bedrock console under Test > Playground, call it through the Anthropic Messages API via bedrock-runtime, or use the Invoke and Converse APIs through the AWS CLI and AWS SDK. AWS also says it’s available today on Bedrock through the Global CRIS inference profile, and through Claude Platform on AWS in North America.
My take — AI-written commentary, not fact-checked reporting
This is the sensible AI strategy everyone keeps talking around: put the cheap model on the repetitive stuff and save the fancy one for the mess. The real story is not another shiny benchmark, but AWS trying to make AI procurement look like normal cloud ops, with guardrails, logs, and a bill line item. Very unglamorous. Very likely to sell.
Read more about this at: Amazon Web Services