Key facts
- Anthropic released Claude Sonnet 5.5 on Monday, an upgrade to its mid-tier AI model.
- The company claims Sonnet 5.5 is over 30% faster and uses nearly a third less tokens per task than its predecessor, Sonnet 5.
- Sonnet 5.5 is priced at $2 per million input tokens and $10 per million output tokens, half the cost of Opus 5.5.
- On the Terminal-Bench 4.0 coding test, Sonnet 5.5 scored 70.6%, outperforming Opus 5.5's 66.4%.
- Independent tester Artificial Analysis ranked Sonnet 5.5 second behind Opus 5.5 but noted it used more tokens per task.
Anthropic released Claude Sonnet 5.5 on Monday, an upgrade to its mid-tier AI model that the company claims is over 30% faster and uses nearly a third less tokens per task than its predecessor, Sonnet 5. The model is priced at $2 per million input tokens and $10 per million output tokens, which is half the cost of its higher-tier Opus 5.5 model.
Anthropic states that Sonnet 5.5 excels at well-scoped everyday tasks, bug fixing, and creating polished documents, slides, and spreadsheets. On the Terminal-Bench 4.0 test, which measures an AI agent's ability to complete complex professional tasks, Sonnet 5.5 scored 70.6%, surpassing Opus 5.5's score of 66.4%. Independent testing firm Artificial Analysis corroborated these findings, scoring Sonnet 5.5 at 63.6% and Opus 5.5 at 59.6%.
However, Artificial Analysis noted that Sonnet 5.5 used more tokens per task than any model it tested at its highest effort setting, which could increase costs. Anthropic claims that at medium effort settings, Sonnet 5.5 offers significant savings for coding tasks. OpenAI recently matched Sonnet 5.5's pricing for its GPT-6 Sol model. Anthropic also indicated that its smallest model, Claude Haiku 5.5, is expected in the coming weeks.