Anthropic launches Claude Opus 5.5 at 40% lower cost
Original: Claude Opus 5.5
Why This Matters
A 60% cache-read price cut directly targets the agentic workflow market where token costs compound fast.
Anthropic released Claude Opus 5.5 on September 22, 2026, the first model in its Claude 5.5 family. It matches Claude Fable 5.1 performance on most tasks, costs 40% less than Opus 5 to run, and scores the highest of any Anthropic model on the company's automated behavioral audit.
Claude Opus 5.5 arrives as Anthropic's first major release since the company publicly called for pacing frontier AI development. Pre-release testing involved external evaluators Frontier Design and METR. Performance gains are substantial: one tester completed a 680,000-line code migration in under a day — work projected to take an engineering team weeks. On a web app load-time task, Opus 5.5 succeeded 39 out of 40 times while Opus 5 made smaller improvements that also changed the app's behavior. In a game-building test from a single prompt, Opus 5.5 outscored all other models on graphics and polish.
Pricing drops are significant. Input tokens fall to $4 per million (down 20%) and output to $20 per million (down 20%), while cache reads — the dominant cost in agentic and coding workflows — fall to $0.20 per million, a 60% reduction. Output speed is more than 30% faster than Opus 5. Anthropic is also raising five-hour usage limits across Pro, Max, Team, and seat-based Enterprise plans and offering subscribers a banked rate-limit reset.
On safety, the model is rated the most aligned in Anthropic's history on its automated behavioral audit, with lower rates of hard-to-reverse actions and stronger resistance to prompt injection. Because Opus 5.5 is comparable to Claude Mythos 5.1 in biology and cybersecurity capability, it ships with equivalent access controls: vetted life sciences organizations can apply now; a Cyber Verification Program expansion is coming within weeks. Claude Sonnet 5.5 and Haiku 5.5 are expected to follow shortly.