Anthropic is cutting the time software vendors get to fix security flaws from 90 days to 30, arguing AI can now find — and rediscover — the same bugs far faster than the old clock assumed. The new timelines take effect for vulnerabilities reported on or after January 1, 2027.

Under the updated coordinated vulnerability disclosure policy, announced on Monday, high- and critical-severity issues will be published 30 days after they are reported if no fix has shipped, down from 90. All other issues get 45 days. Reports already in progress when the new rules begin will keep their existing deadlines.

Anthropic is also tightening two related clocks. If a maintainer does not respond to a report, the company will escalate to an external vulnerability coordinator after 15 days instead of 30. Once a patch is available, full technical details will be shared after 30 days rather than 45. The seven-day window for vulnerabilities already under active exploitation is unchanged.

Why the 90-day clock no longer works

Coordinated disclosure norms were written before AI could discover vulnerabilities, Anthropic said. They assumed fixes moved at human speed and that the same bug was unlikely to be found independently by someone else. Neither assumption holds today: models can now help develop fixes, and a bug one model can find is increasingly likely to be found by another. The company says rediscovery is already rising as these capabilities improve.

The change follows a wave of evidence that AI has collapsed the time between discovery and exploitation. Earlier on Monday, security firm Forescout reported that AI models found 66 previously unknown flaws in critical infrastructure code in just 90 hours — work its researchers said would have taken humans about nine months. Anthropic's own tools are part of that shift: its Claude Mythos Preview model can reverse-engineer exploits on closed-source software, and its free OSS Scanner has already surfaced more than 29,000 candidate vulnerabilities in open-source projects. The 90-day standard itself comes from Google's Project Zero, the industry benchmark Anthropic is now breaking from.

What happens to overdue reports

The new policy also clears a backlog. Anthropic said there are high- and critical-severity vulnerabilities it reported more than 90 days ago that still have not been disclosed publicly, because the maintainers have not fixed them. Those projects are now being told they have an additional 30 days, after which the reports will be published on Anthropic's dashboard starting January 1, 2027.

Exceptions and help for maintainers

Not every flaw will face the shorter clock. Bugs in hard-to-deploy software, such as critical infrastructure, and flaws that require changes to a specification or RFC cannot always be resolved in 30 or 45 days, Anthropic said — in those cases it will agree a remediation timeline with the vendor. The new policy also does not apply to AI-discovered findings that have not received human review, such as raw OSS Scanner output.

Anthropic acknowledged that shorter timelines put more pressure on software maintainers, and said it will keep including actionable guidance with each report, along with a patch or mitigation where possible, and will agree a sustainable pace before sending a large volume of findings to a single project. It also pointed to programs designed to help defenders keep up: OSS Scanner scans critical open-source projects free of charge, and the Claude for Open Source Program offers six months of Claude Ultra at no cost.

The disclosure clock was built for a world where finding a bug took months. In a world where AI finds them in hours, Anthropic is betting 30 days is the new normal.