Anthropic launches Claude Sonnet 5.5 with promised speed and cost gains
Anthropic says it runs more than 30% faster than Sonnet 5 and costs up to 30% less for most work. In pre-release tests, Artificial Analysis found the max-effort setting cost roughly 50% more per task than Sonnet 5.
TLDR
Anthropic launched Claude Sonnet 5.5, claiming it runs more than 30% faster than Sonnet 5 and costs up to 30% less for most work. GitHub says the model is generally available in Copilot and matched Sonnet 5 on coding tasks in its early tests while using fewer steps, tokens and tool calls. Artificial Analysis ranked Sonnet 5.5 at max effort No. 2 on its Intelligence Index, but measured a per-task cost roughly 50% higher than Sonnet 5’s. Those tests used a pre-release version with a structured-output bug that Anthropic says it fixed for the public release; Artificial Analysis plans to rerun relevant evaluations.
Combined views
1.6M
60 Sources, first seen 7h ago
