Anthropic released Claude Opus 5.5, the first model in its new Claude 5.5 family, claiming the top spot on agentic coding and knowledge work benchmarks while costing 40% less to run than Opus 5. Pricing drops to $4 per million input tokens and $20 per million output tokens, 20% below its predecessor, with cache reads down 60% to $0.20 per million. Output generates more than 30% faster, and Anthropic is raising five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans while giving subscribers a saveable rate-limit reset.
The company says Opus 5.5 matches Claude Fable 5.1 on most work. It leads Anthropic's own Terminal-Bench 4.0 agentic coding test at 66.4% against Fable 5.1's 55.8%, and takes 54.4% on FrontierCode v1.1 versus 50.3%. One early tester completed a 680,000-line code migration in under a day. Another asked it to cut load times across an entire web app, where it succeeded 39 times out of 40. On the automated behavioral audit Anthropic calls its most comprehensive alignment test, Opus 5.5 scores the best of any model the company has tested.
Anthropic frames the launch as its first since calling for pacing the frontier, with external evaluation by METR and Frontier Design. Because biology and cybersecurity capability approaches Claude Mythos 5.1, access ships with the same safeguards as Fable 5.1: vetted labs apply through the Life Sciences Verification Program, and a Cyber Verification Program expands in coming weeks. Sonnet 5.5 and Haiku 5.5 follow shortly. The announcement is the top story on Hacker News with over 900 points.