Anthropic's Claude Opus 5.5 is faster, cheaper and matches Fable 5.1 on most work
Anthropic has launched Claude Opus 5.5 , the first model in its new Claude 5.5 family, and the biggest change is cost. The company says the new AI model performs at the level of Claude Fable 5.1 on most work while costing 40% less to run than Opus 5. It also generates output more than 30% faster than its predecessor.
This is Anthropic's first release since it called for pacing the AI frontier, so safety features sit right next to performance. Opus 5.5 was tested before launch by external evaluators, including METR and Frontier Design. Anthropic says it scored the best of any model on its automated behavioural audit, which runs nearly 2,000 simulated scenarios. The model is available now on the Claude Platform as claude-opus-5-5, and on Amazon Web Services, Google Cloud and Microsoft Azure.
Claude Opus 5.5 pricing undercuts Opus 5 on every tokenInput tokens cost $4 (about Rs 350) per million and output tokens $20 (about Rs 1,760), both 20% lower than Opus 5. Cache reads, which Anthropic says make up most of the bill in agentic and coding work, drop 60% to $0.20 (about Rs 18) per million. Opus 5.5 also uses fewer tokens per task, which is where the overall 40% saving comes from.
Subscribers benefit too. Anthropic is raising five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans. Users also get a rate limit reset they can save and use whenever they choose.
Opus 5.5 leads benchmarks in coding and knowledge workOn Terminal-Bench 4.0, Opus 5.5 scored 66.4%, ahead of Fable 5.1 at 55.8% and OpenAI's GPT-6 Astra at 57.9%. On GDPval-AA v2.1, which tests real-world work across 44 occupations, it posted 1846 Elo against 1708 for Opus 5.
Anthropic also shared results from actual workloads. One early tester completed a 680,000-line code migration in under a day. In an internal test, Opus 5.5 and Fable 5.1 both rewrote HAProxy, a widely used load balancer, from C into Rust. Opus 5.5 finished in 9.5 hours against 12 and cost 51% less.
The company says the model writes more clearly as well. It puts key information first and avoids jargon, which responds to a common complaint about Opus 5.
Anthropic adds Fable-level safeguards for biology and cybersecurityOpus 5.5 matches Claude Mythos 5.1 in biology and cybersecurity, so it ships with safeguards similar to those on Fable 5.1. Developers can still find and fix bugs in their own code, but most cybersecurity tasks will be rerouted to Opus 4.8. Vetted labs, startups and pharmaceutical firms can apply to a new Life Sciences Verification Program . Verified security practitioners will get access through an expanded Cyber Verification Program in the coming weeks.
Anthropic acknowledges some limits. It says Opus 5.5 often suspects it is being evaluated, which makes its real-world behaviour harder to predict. The company is pairing its alignment work with these safeguards for that reason.
Claude Sonnet 5.5 and Claude Haiku 5.5 are due in the coming weeks. For now, Anthropic's cheapest Opus in a generation is also the one it rates safest, even though its own tests may not catch everything.
This is Anthropic's first release since it called for pacing the AI frontier, so safety features sit right next to performance. Opus 5.5 was tested before launch by external evaluators, including METR and Frontier Design. Anthropic says it scored the best of any model on its automated behavioural audit, which runs nearly 2,000 simulated scenarios. The model is available now on the Claude Platform as claude-opus-5-5, and on Amazon Web Services, Google Cloud and Microsoft Azure.
Claude Opus 5.5 pricing undercuts Opus 5 on every tokenInput tokens cost $4 (about Rs 350) per million and output tokens $20 (about Rs 1,760), both 20% lower than Opus 5. Cache reads, which Anthropic says make up most of the bill in agentic and coding work, drop 60% to $0.20 (about Rs 18) per million. Opus 5.5 also uses fewer tokens per task, which is where the overall 40% saving comes from.
Subscribers benefit too. Anthropic is raising five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans. Users also get a rate limit reset they can save and use whenever they choose.
Opus 5.5 leads benchmarks in coding and knowledge workOn Terminal-Bench 4.0, Opus 5.5 scored 66.4%, ahead of Fable 5.1 at 55.8% and OpenAI's GPT-6 Astra at 57.9%. On GDPval-AA v2.1, which tests real-world work across 44 occupations, it posted 1846 Elo against 1708 for Opus 5.
Anthropic also shared results from actual workloads. One early tester completed a 680,000-line code migration in under a day. In an internal test, Opus 5.5 and Fable 5.1 both rewrote HAProxy, a widely used load balancer, from C into Rust. Opus 5.5 finished in 9.5 hours against 12 and cost 51% less.
The company says the model writes more clearly as well. It puts key information first and avoids jargon, which responds to a common complaint about Opus 5.
Anthropic adds Fable-level safeguards for biology and cybersecurityOpus 5.5 matches Claude Mythos 5.1 in biology and cybersecurity, so it ships with safeguards similar to those on Fable 5.1. Developers can still find and fix bugs in their own code, but most cybersecurity tasks will be rerouted to Opus 4.8. Vetted labs, startups and pharmaceutical firms can apply to a new Life Sciences Verification Program . Verified security practitioners will get access through an expanded Cyber Verification Program in the coming weeks.
Anthropic acknowledges some limits. It says Opus 5.5 often suspects it is being evaluated, which makes its real-world behaviour harder to predict. The company is pairing its alignment work with these safeguards for that reason.
Claude Sonnet 5.5 and Claude Haiku 5.5 are due in the coming weeks. For now, Anthropic's cheapest Opus in a generation is also the one it rates safest, even though its own tests may not catch everything.
Next Story