Anthropic Cuts Opus 5.5 Price to $20, Tops Fable Benchmarks

Anthropic released Opus 5.5 on Tuesday, and the company says its newest flagship model beat the larger Fable system on multiple benchmarks while cutting the price developers pay to run it, according to TechCrunch. The release lands just two months after Opus 5 launched on July 24, keeping Anthropic on a fast cadence in its three-tier Claude lineup of Opus, Sonnet and Haiku.

What Did Anthropic Announce With Opus 5.5?

Opus is Anthropic's most capable and most expensive tier, sitting above Sonnet and Haiku. The company says Opus 5.5 sets a new state-of-the-art mark in coding and knowledge-work performance and outpaces Fable — a larger Anthropic model — on many benchmarks. Anthropic also says Opus 5.5 succeeded at a number of informal tasks that Fable could not complete, though the company did not specify which tasks in the material reviewed. The model runs faster than its predecessor, which Anthropic attributes to an overall drop in the compute required to serve it. Anthropic also revised how Opus communicates, making it less likely to lean on jargon and more likely to put the most important information at the top of a response.

How Much Cheaper Is Opus 5.5 Than the Previous Version?

Output tokens for Opus 5.5 will be billed at $20 per million tokens, down from $25 per million for Opus 5, per the TechCrunch report. Other pricing metrics saw comparable reductions. The cut arrives alongside the performance gains rather than as a trade-off, according to Anthropic, since the faster runtime reflects lower compute costs on the company's end.

What Safety Rules Apply to the New Model?

Anthropic says Opus 5.5's biology and cybersecurity capabilities are comparable to Mythos, its most advanced model, which means Opus 5.5 falls under the same safeguards. Those restrictions limit how the model can be used to find exploits in compiled programs or to assist in developing recognizable biological weapons, according to the company.

The release is Anthropic's first since CEO Dario Amodei publicly embraced calls to slow the pace of capability advances relative to safety work. In a post cited by TechCrunch earlier this month, Amodei wrote:

"I have become convinced that fully addressing the risks requires even more prudence, not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up."

Anthropic says Opus 5.5's safety training followed a similar process to prior releases, including alignment testing and pre-release evaluation by outside groups METR and Frontier Design. The company says more advanced training and evaluation systems, including improved security and monitoring, are already being prepared for future models. "As AI becomes more capable, public policy should play a larger role in making sure the systems people rely on are safe," Anthropic wrote in its release announcement, adding that it has "started to put the infrastructure in place to support it" and plans to "share more details on these efforts soon."

When Will Sonnet and Haiku Get the Same Upgrade?

Anthropic says Sonnet 5.5 and Haiku 5.5 — the next two tiers down — are coming "in the coming weeks" with performance improvements similar to what Opus 5.5 delivers. No specific release date was given for either model.

Timeline

• July 24, 2026 — Opus 5 launches.
• September 2026 — Amodei publishes post calling for pacing capability growth to match safety work.
• September 22, 2026 — Opus 5.5 launches with lower pricing and benchmark claims against Fable.
• Coming weeks — Sonnet 5.5 and Haiku 5.5 expected.

What to Watch

• Whether independent evaluators publish benchmark comparisons between Opus 5.5 and Fable beyond Anthropic's own claims.
• The release dates and pricing for Sonnet 5.5 and Haiku 5.5.
• Further detail from Anthropic on the "infrastructure" and public-policy role it says it is building for future model oversight.
• Whether rival labs respond with their own price cuts following Anthropic's move to $20 per million output tokens.

Anthropic did not disclose specific benchmark scores or task names in the material reviewed, and the company's comparisons to Fable and Mythos have not been independently verified.