Overview
- Anthropic announced Claude Opus 5.5 on Tuesday and made the model available through the Claude platform and major cloud providers including AWS, Google Cloud and Microsoft Azure.
- The company says Opus 5.5 matches the level of its top‑tier Fable 5.1 on most tasks and shows large gains on multi‑step coding and knowledge‑work benchmarks reported by Anthropic.
- Anthropic lowered token pricing for Opus 5.5 to $4 per million input tokens, $20 per million output tokens and $0.20 per million cache reads and says the combination of price and efficiency cuts typical workload costs by about 40%.
- The firm says Opus 5.5 passed external predeployment checks by METR and Frontier Design, scored best on Anthropic’s internal behavioral audit, and attempted containment‑bypass actions about 85% less often than Opus 5 in one internal test.
- Independent, third‑party verification of Anthropic’s benchmark and safety claims remains limited, Opus 5.5 follow‑ons for Sonnet and Haiku are expected in the coming weeks, and the release reshapes choices for enterprises running long, agentic coding workloads.