Anthropic has released Claude Opus 5.5, the first model in its Claude 5.5 family. The company says it performs at the level of Claude Fable 5.1 on most work and, at default settings on typical workloads, costs 40% less to run than Opus 5.

The timing is notable: as TechCrunch reported, Anthropic published the model just 90 minutes before OpenAI announced GPT-6 Sol and Luna. Both companies made the same promise in the same week: a little more capability for noticeably less money.

Where the saving comes from

Anthropic says Opus 5.5 needs less compute to serve than Opus 5, and the pricing reflects that. The gain comes from three places:

  • Cache reads, which make up most of the cost in agentic and coding work, are 60% cheaper.
  • The model spends fewer tokens per task.
  • Output generation is more than 30% faster than Opus 5.

Together these net out to the 40% cost reduction on typical workloads. Fast mode in Claude Code and the Claude Platform offers up to 2.5x the speed at $8 input and $40 output per million tokens. Anthropic is also raising five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans.

Benchmarks: a lead, not a clean sweep

On Anthropic's own tests, Opus 5.5 leads in agentic coding, computer use and knowledge work. But the company adds two caveats of its own: GPT-6 Astra still leads on Terminal-Bench-Science and AutomationBench, and benchmark margins are becoming a less reliable guide. In its own use, Anthropic says, the gap to Fable 5.1 is narrower than the scores suggest.

The cost-adjusted results say more. At default (medium) effort, Opus 5.5 scores 54.6% on FrontierCode, beating GPT-6 Astra's top score of 53.3% at about a fifth of the cost per task. On CursorBench at medium effort it scores 52.5%, eleven points above GPT-5.6 Sol's best result and at roughly a third of the cost.

What early testers reported

The examples Anthropic published are about long-running work:

  • One tester completed a 680,000-line code migration in less than a day.
  • Another audited and fixed a 200,000-line codebase in under three hours; the same job took Opus 5 more than 20 hours and 2.5 times the tokens.
  • In an internal C-to-Rust port, Opus 5.5 finished in 9.5 hours against 12 hours for Fable 5.1, and cost 51% less.
  • Deloitte reports that Opus 5.5 at its lowest effort caught 72% of known review bugs, against 56% for Opus 5 at high effort.

All of these come from the vendor and from testers it selected.

Access

The model can be called as claude-opus-5-5 on the Claude Platform, Amazon Web Services, Google Cloud and Microsoft Azure. Zero data retention is available, as with previous Opus models. No weights were released, so self-hosting is not an option.

For developers the real question is this: when two launches make the same promise in the same week, no vendor chart can tell you which one is cheaper for your work. Only measuring it on your own workload can.