Anthropic has released Claude Sonnet 5.5 for coding, documents, and other well-defined work.
Anthropic has released Claude Sonnet 5.5 for coding, documents, and other well-defined work. Its API list price remains $2 per million input tokens and $10 per million output tokens, the same as Sonnet 5. Anthropic says output is more than 30% faster and its own tests show up to 30% less cost per completed task because the model uses fewer tokens. Those are the company's measurements, not a saving every application should expect.
The model ID is claude-sonnet-5-5. Claude Platform's release notes list availability through Anthropic's API, Amazon Bedrock, Claude Platform on AWS, Google Cloud, and Microsoft Foundry. Claude Code 2.1.284 also makes it the default Sonnet model.
Per-token pricing tells only part of the cost story. An agent that needs fewer tool calls or retries may cost less to complete a job even when the published token rate stays fixed. VentureBeat asked Anthropic for more detail on its speed measurement. Teams paying for API usage should compare representative tasks for total cost, latency, and output quality before changing their default model.
There is also a compatibility check. Anthropic's migration notes warn that some requests written for Sonnet 5 can fail on Sonnet 5.5. Forced tool selection with tool_choice: any or tool returns HTTP 400. The older computer_20251124 computer-use tool is not accepted on Anthropic's API or Google Cloud. Applications using either feature need a request-level migration, rather than only a new model name.
In Anthropic's own Terminal-Bench 4.0 evaluation, Sonnet 5.5 scored 70.6%, compared with 10.3% for Sonnet 5. That result describes one benchmark under its stated conditions; it does not establish the same gain on another team's code. Anthropic still positions Opus 5.5 for complex, open-ended work that needs sustained judgment.