Introducing Claude Opus 5.5

Open original

Anthropic launched Claude Opus 5.5, a new flagship model that matches the performance of their previous Fable 5.1 model while costing 40% less to run. The model shows improvements in safety testing, coding ability, and communication clarity, with faster output generation and lower per-token pricing.

  • Pricing drops 40% overall: input tokens at $4/million (down from $5), output at $20/million (down from $25), and cache reads at $0.20/million (down from $0.50)—a 60% reduction on the cache reads that dominate coding workloads.
  • Performance gains on real tasks: one tester completed a 680,000-line code migration in under a day that would have taken weeks; another used it to audit and fix a 200,000-line codebase in under three hours versus Opus 5's 20+ hours.
  • Safety scores higher on Anthropic's behavioral audit than any previous model tested, with stronger resistance to prompt injection attacks and improved ability to stay within given boundaries.
  • Output generation is 30% faster than Opus 5, with a "fast mode" option available at 2.5x speed for additional cost.

Anthropic released Claude Opus 5.5, a more efficient version of their flagship model that matches Claude Fable 5.1's performance while costing 40% less to run. The pricing drops significantly: input tokens cost $4 per million (down from $5), output tokens $20 (down from $25), and cached reads just $0.20 (down from $0.50)—a 60% reduction on cache operations that dominate agentic and coding workloads. The model also runs 30% faster than Opus 5. Anthropic tested it with external evaluators from Frontier Design and METR before release and claims it scores highest on their behavioral alignment audit, showing better resistance to prompt injection and less likelihood to take hard-to-reverse actions. The performance gains are real but incremental on benchmarks. Opus 5.5 leads on agentic coding, computer use, and knowledge work tasks, but Anthropic acknowledges that benchmark margins have become unreliable guides to actual differences. Real-world testing reveals the practical advantage: one early user completed a 680,000-line code migration in under a day versus weeks for a team, and in web app optimization, Opus 5.5 succeeded 39 of 40 times while Opus 5 made smaller improvements that broke functionality. On a HAProxy translation from C to Rust, Opus 5.5 finished in 9.5 hours versus 12 for Fable 5.1 at 51% lower cost. The model communicates more naturally than previous versions, putting important information first and writing in a clearer style that testers found easier to follow—one described it as writing "the way I do." For sensitive applications, Anthropic deployed it with restricted access: biology research requires enrollment in their Life Sciences Verification Program, and cybersecurity work needs verification through their Cyber Verification Program. Claude Sonnet 5.5 and Haiku 5.5 are coming in the following weeks with similar improvements to performance, efficiency, and safety.

Ask about this article