Anthropic has released Claude Sonnet 5.5, which ranks second on the Artificial Analysis Intelligence Index, an independent benchmark for AI models. [1, 4]
Sonnet 5.5 scored 56. [4] OpenAI's GPT-6 Astra scored 53, while Anthropic's Opus 5.5 led with 58. [4] Sonnet 5.5 improved by 18 points over its predecessor, Sonnet 5. [4]
To achieve its top ranking, Sonnet 5.5 uses approximately seven times as many output tokens as GPT-6 Astra. [1, 4] Anthropic kept the list price unchanged, but the higher token use raises the cost per task for workloads seeking maximum performance. [1, 13]
Sonnet 5.5 is available on AWS, Google Cloud, and Microsoft Azure. [3]