Claude Sonnet 5.5 is Faster, Smarter and Up to 30% Cheaper

Claude Sonnet 5.5, Anthropic's new mid-tier AI model, offers significant improvements in speed, intelligence, and cost-effectiveness. The model excels in coding tasks, scoring 70.6% on Terminal-Bench 4.0, compared to 10.3% for Sonnet 5. It also performs well on FrontierCode 1.1 and CursorBench 4.0. Claude Sonnet 5.5 can process large codebases quickly and complete coding tasks in fewer steps by optimizing tool calls.

The model handles tens of thousands of lines of code and can sustain multi-hour coding tasks while maintaining fast response times. In everyday tasks like bug fixing, document creation, presentations, and spreadsheet work, Sonnet 5.5 is designed for speed, capability, and affordability. It improves computer-use performance and visual chart recognition, scoring 80.1% and 61.6% respectively on Anthropic's evaluations.

The model also demonstrates stronger long-horizon reasoning and image understanding, including beating Pokémon Red from screenshots, which highlights its ability to interpret visual information over extended sequences. Despite having the same token prices as Sonnet 5, Claude Sonnet 5.5 generally requires fewer tokens to complete tasks, making it up to 30% cheaper per task.

Additionally, Sonnet 5.5 introduces stronger safeguards, including cybersecurity improvements, which make it the first Sonnet model to launch with protections similar to those used for Anthropic's more capable models. Claude Sonnet 5.5 is now available on Anthropic's platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure.