Anthropic has introduced Claude Sonnet 5.5, a model aimed at coding, research and the everyday work that fills a busy afternoon. Anthropic says it generates output more than 30% faster than Sonnet 5, while keeping the same API list price of $2 per million input tokens and $10 per million output tokens. It supports a context window of up to one million tokens.
The most striking gains appear in Anthropic’s published evaluations:
| Benchmark | Sonnet 5 | Sonnet 5.5 | Opus 5.5 |
|---|---|---|---|
| Terminal Bench 4.0 | 10.3% | 70.6% | 66.4% |
| CursorBench 4.0 | 34.1% | 55.5% | 57.8% |
| OSWorld 2.1 | 57.0% | 80.1% | 81.8% |
| Humanity’s Last Exam, with tools | 54.9% | 64.5% | 67.7% |
These are Anthropic’s results, and scores can change with the test setup and reasoning effort. They point to a Sonnet model that can handle substantially more demanding tasks, particularly in coding and computer use.
The unchanged token price is welcome, but it does not guarantee that every task costs the same. Artificial Analysis found that Sonnet 5.5 uses considerably more tokens at its maximum effort setting. The right choice depends on the quality, speed and total cost of the work you actually run.
Read Anthropic’s Claude Sonnet 5.5 announcement, then explore what the latest AI models can do at opencheese.dev.