Anthropic's Cheaper AI Helper Now Lives Inside Amazon's Cloud

Anthropic has put a new version of its Claude AI, called Sonnet 5.5, inside Amazon's cloud platform, AWS. If your company already uses Amazon for its software, this is a bit like a new employee showing up who's already badged into the building. Nothing extra to install, nothing new to buy — it just appears in the tools your tech team already uses. Amazon handles the security, logging and spending controls, and it all shows up on the same monthly Amazon bill.

The headline upgrade is cost and speed, not raw brilliance. Anthropic has two models: Opus 5.5, the expensive one for hard judgment calls like untangling messy code or reviewing contracts, and Sonnet 5.5, the cheaper, quicker one for work where the plan is already clear. Think of it as the difference between hiring a senior consultant for a problem you don't understand and having a fast, reliable assistant handle the checklist. Sonnet 5.5 is meant for the checklist — over and over, all day, without costing a fortune.

In practice, that means tasks like being the first responder when a system sends an alarm at 2am, generating routine database queries, writing short documents and one-page summaries, making architecture diagrams and slide decks, and editing spreadsheets in small ways for hundreds of employees at once. Anthropic also says it's better at checking its own work against what was actually asked — a real problem with AI, which often produces something confident and slightly wrong.

The catch is that "cheaper per task" only helps if you trust the output enough to skip reviewing it. The model is designed for well-scoped jobs with a clear finish line. Hand it something vague or high-stakes and you'll still want the expensive model or a human. Also worth noting: this is an enterprise product, so you can't sign up as an individual — your company's tech team has to switch it on first.