Claude Opus 5.5 Costs Less Than Opus 5 and Leads Anthropic’s Coding Benchmarks

Anthropic's Claude Opus 5.5 is 20% cheaper per token than Opus 5 and tops the company's coding and knowledge-work benchmarks. Developers moving over will need to handle a few breaking API changes.
Anthropic announcement artwork with the words Claude Opus 5.5 over a horizon at dusk
Claude Opus 5.5 Costs Less Than Opus 5 and Leads Anthropic’s Coding Benchmarks

Updated Oct. 24, 2026, 3:11 p.m. ET: Adds the releases of Claude Sonnet 5.5 and Claude Haiku 5.5.

Anthropic’s newest flagship costs less than the model it replaces. Claude Opus 5.5, released Tuesday as the first model in the Claude 5.5 family, is priced at $4 per million input tokens and $20 per million output tokens, 20% below Claude Opus 5.

The bigger saving is in cache reads, which Anthropic said make up most of the cost of agentic and coding work. Those dropped 60%, to $0.20 per million tokens. Because the model also uses fewer tokens per task, the company said typical workloads at default settings should cost about 40% less than on Opus 5, and it said output comes back more than 30% faster.

Benchmarks and early testers

On Anthropic’s own tables, Opus 5.5 scored 66.4% on Terminal-Bench 4.0, ahead of Opus 5 at 52.3% and OpenAI’s GPT-6 Astra at 57.9%, and it posted 1846 Elo on GDPval-AA v2.1, a test of professional work across 44 occupations. The company added a caveat of its own: “at these levels of capability we’ve found that benchmark margins have become a less reliable guide to real-world differences.”

Testers described the gains in hours saved. One audited and fixed a 200,000-line codebase in under three hours, a job Opus 5 needed more than 20 hours for, according to Anthropic. GitHub Chief Product Officer Mario Rodriguez said that in VS Code the model “solved more terminal tasks than Opus 5 in less than half the steps.”

Anthropic also said it worked on how the model writes, after feedback about Opus 5’s output. “It puts the most important information up front,” the company said.

Safeguards and changes for developers

Outside evaluators, including METR and Frontier Design, tested Opus 5.5 before release, and Anthropic said it tried to cross containment boundaries about 85% less often than Opus 5 in a new evaluation. Because its biology and cybersecurity skills rival Claude Mythos 5.1, most cybersecurity tasks are routed to Claude Opus 4.8. Vetted labs and companies can apply to a new Life Sciences Verification Program for fuller biology access, TechRepublic reported.

Existing apps may need work before switching. Per Anthropic’s developer notes, thinking can’t be disabled on Opus 5.5, forced tool use now returns an error, and the older computer_20251124 computer use tool isn’t accepted on the Claude API.

The model is available through the Claude API as claude-opus-5-5 and on Amazon Bedrock, Google Cloud and Microsoft Foundry, with a 1 million-token context window. Google Cloud customers later got the option to have Anthropic run Claude directly. Anthropic also raised five-hour usage limits on its Pro, Max, Team and seat-based Enterprise plans.

The rest of the family has since followed: Claude Sonnet 5.5 arrived Sept. 28 at $2 and $10 per million input and output tokens, and Claude Haiku 5.5 launched Oct. 7, according to Anthropic’s release notes.

Subscribe for Updates

GenAIPro Newsletter

News, updates and trends in generative AI for the Tech and AI leaders and architects.

By signing up for our newsletter you agree to receive content related to ientry.com / webpronews.com and our affiliate partners. For additional information refer to our terms of service.

Notice an error?

Help us improve our content by reporting any issues you find.

Get the WebProNews newsletter delivered to your inbox

Get the free daily newsletter read by decision makers

Subscribe
Advertise with Us

Ready to get started?

Get our media kit

Advertise with Us