Anthropic Releases Claude Opus 5.5: 40% Cheaper, 30% Faster, Safer.

Anthropic has launched Claude Opus 5.5, the first model in the new Claude 5.5 family, delivering performance comparable to Claude Fable 5.1 across most workloads while reducing operational costs by 40% relative to Opus 5. Released under Anthropic's pacing the frontier framework, the model emphasizes alignment and safety alongside efficiency. Opus 5.5 demonstrates significant improvements in agentic coding and knowledge work. Early evaluations show the model completing a 680,000-line code migration in under a day, whereas previous versions required weeks. It achieved a 39 out of 40 success rate in optimizing web application load times without altering functionality. On benchmarks, Opus 5.5 leads Terminal-Bench 4.0, FrontierCode v1.1, and CursorBench 4.0. In knowledge tasks, it scored 1,846 Elo on GDPval-AA v2.1, outperforming competitors. Financial analysis tests revealed a 32% speed increase and 50% cost reduction compared to Opus 5. Cost efficiency is a defining feature. Pricing drops to $4 per million input tokens, $20 per million output tokens, and $0.20 per million cache reads, with output generation speeds improving by over 30%. Anthropic reports that at default settings, Opus 5.5 matches GPT-6 Astra on coding benchmarks at roughly 20% of the cost. Subscription users on Pro, Max, Team, and Enterprise plans now benefit from increased five-hour usage limits and rate limit resets. Safety protocols mirror those applied to Claude Fable 5.1. Opus 5.5 achieved record scores on Anthropic's automated behavioral audit, showing heightened resistance to prompt injection and reduced attempts to bypass containment. Pre-release testing involved external evaluators Frontier Design and METR. For high-risk domains, access is restricted. Most cybersecurity tasks are routed to Claude Opus 4.8, while biology research requires enrollment in the Life Sciences Verification Program. An expanding Cyber Verification Program will grant vetted practitioners access. The model also includes preserved thinking to mitigate distillation attacks. Users note improved communication, with Opus 5.5 providing clearer, more direct responses that prioritize critical information. The model is available on Amazon Web Services, Google Cloud, Microsoft Azure, and the Claude Platform via the identifier claude-opus-5-5. Claude Sonnet 5.5 and Claude Haiku 5.5 are scheduled for release in the coming weeks.