Claude Sonnet 5.5 is out, and Anthropic says it does the same jobs faster for less

Anthropic released Claude Sonnet 5.5 on 28 September. The per-token price hasn't changed, but Anthropic says it finishes most work over 30% faster and at up to 30% lower cost per task.

Anthropic released Claude Sonnet 5.5 on 28 September. Sonnet is the middle model in the Claude line-up, the one most people use for everyday writing, analysis and coding, and this version replaces Sonnet 5.

The price per token is the same as before: $2 per million input tokens and $10 per million output tokens. The saving comes from the model using fewer tokens to finish a job. Anthropic says Sonnet 5.5 runs more than 30% faster than Sonnet 5 and costs up to 30% less per task. Some of the customer results it published point the same way: Zendesk reported 20% faster ticket processing, and Box said it was 2.4 times faster with better accuracy. It is available now in the Claude apps on web, desktop and mobile, through Anthropic's own platform, and on AWS, Google Cloud and Microsoft Azure.

For a midsize business, the change you will feel first is speed in the apps your team already uses. If you pay for Claude through the API, or through a tool built on top of it, the lower cost per task matters more. A support inbox, a document review process or a reporting job that runs hundreds of times a week should cost less to run without anyone changing how it works.

Those figures are Anthropic's and its customers', measured on their own work. Before you switch a production process over, run the new model on ten or twenty real examples from your business next to the old one and compare the output and the bill. Check with any vendor whose product runs on Claude whether and when they plan to move to 5.5.

A faster, cheaper version of the model you already use is the easiest kind of upgrade to take. Test it on your own work first, then move over.