Anthropic launches Claude Opus 5.5, cutting costs by 40%
The new model matches a higher-tier competitor on benchmarks while dropping API prices and tightening safety guardrails
Anthropic’s latest model arrives with a rare combination in AI releases: it costs less, runs faster, and is harder to manipulate than the version it replaces. Claude Opus 5.5 launched on September 22, 2026, and the headline number is a 40% reduction in overall operating costs compared to Claude Opus 5, which itself only shipped two months ago.
What changed and what it costs now
The pricing structure tells the clearest story. Input tokens now run at $4 per million, output at $20 per million, each representing a 20% cut from Opus 5’s $5 and $25 rates. Cache reads dropped even further, falling 60% to $0.20 per million tokens.
The model also generates output more than 30% faster than its predecessor, which matters less for casual queries and a great deal for agentic tasks that chain dozens of model calls together. Anthropic built Opus 5.5 specifically for long-duration agentic coding and professional workloads, giving it a 1M-token context window and a maximum synchronous output size of 128K tokens.
On benchmarks, the model matches or exceeds Claude Fable 5.1, Anthropic’s higher-tier offering, across a range of coding and knowledge-work evaluations. On GDPval-AA v2.1, it scores 1,846 Elo.
Safety improvements that are actually measurable
Independent evaluators, including Frontier Design and METR, found that boundary-circumvention attempts fell 85% compared to both Opus 5 and Mythos 5.1. Those are external assessments, not self-reported figures. Anthropic is billing Opus 5.5 as its strongest-performing model in comprehensive alignment testing to date.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
CEO Dario Amodei has been publicly pushing for a more measured pace of capability releases given ongoing safety concerns across the industry, and the emphasis on alignment in this launch reflects that position. Anthropic says the model now prioritizes essential information and reduces jargon.
Who gets access and through what channels
Opus 5.5 is available immediately through the Claude API, the Claude Platform, Amazon Bedrock, and Google Cloud. Microsoft is also listed as a distribution partner, giving the model coverage across the three dominant enterprise cloud environments simultaneously at launch.
Anthropic expanded usage limits alongside the release. Pro, Max, Team, and Enterprise subscribers receive increased five-hour usage caps, with optional rate-limit resets available.
Early user reports highlight code migration tasks as a standout use case. Large-scale refactoring or framework migrations are completing with fewer interruptions, aligning with the model’s design intent for professional developer tooling.