Anthropic raises five-hour usage limits by 20% on Pro, Max and Team plans

Anthropic released Claude Opus 5.5 on September 22, 2026, and attached a change that matters to people already paying for Claude: five-hour usage limits are going up by 20% on the Pro, Max and Team plans. Techmeme's summary of the report by David Gewirtz for ZDNET also notes that subscription users receive a rate limit reset.

Cheaper to run, faster to answer

Less than two months after Claude Opus 5, Anthropic says its new model delivers Fable 5.1-level performance for most work at roughly 40% lower running cost. Opus 5.5 generates output more than 30% faster than Opus 5, and its tokens are priced 20% lower; the model also needs fewer tokens for higher-quality work. Claude Sonnet 5.5 and Haiku 5.5 are due in the coming weeks.

Early customers report similar gains. Yashodha Bhavnani, VP of AI Products at Box, says Opus 5.5 used a third of the tokens Opus 5 did in the company's evaluations, with answers 40% less verbose and no loss of accuracy.

More headroom before the limits bite

The subscription changes reach beyond the five-hour window. Because Opus 5.5 costs less to serve than Opus 5, both the five-hour and weekly limits stretch further, and Gewirtz describes the effects as additive: a 20% larger bucket combined with a 25% slower burn, which works out to roughly 50% more usable capacity. For users on a $20/month plan, he writes, that is a considerable win.

Safeguards and fallbacks

Anthropic presents Opus 5.5 as its better-behaved release, citing CEO Dario Amodei's essay on pacing the frontier. The company says the model went through extensive alignment testing, pre-release evaluation by outside organizations, and safeguards for high-risk areas such as cybersecurity and biology. When those safeguards fire, requests fall back from Opus 5.5 to Opus 4.8: most cybersecurity tasks are rerouted there, while biology and LLM-development requests go to Opus 5. Users approved through Anthropic's Cyber Verification Program will get Opus 5.5 access in a few weeks.

Quality claims extend to code review. Carl Bennett, CIO at Deloitte Consulting, says that even at its lowest effort setting, Opus 5.5 caught 72% of known bugs in the firm's code reviews, against 56% for Opus 5 at high effort, with fewer false alarms and a fraction of the output.