anthropic.com web signal

Anthropic Redeploys Fable 5 With Cross-Lab Jailbreak Rubric

7 sources tracking this story
anthropic amazon microsoft google safety ai-business

TL;DR

  • Commerce conditions bind Anthropic to proactive security monitoring and mandatory malicious-use reporting, creating the first federal AI redeployment compliance framework.
  • Anthropic's four-axis jailbreak severity framework is the first shared triage standard proposed across Amazon, Microsoft, and Google.
  • Katie Moussouris, the only outside reviewer of the underlying research, called the flagged behavior routine defensive security work, not a guardrail bypass.

Anthropic's redeployment note for Fable 5 does double duty. On the surface it is a return to service: the model comes back to Claude Platform, Claude.ai, Claude Code, and Claude Cowork on July 1, after the US government lifted export controls that had been in place since June 12. Underneath it is the closest thing the frontier labs have offered so far to a shared vocabulary for jailbreaks.

The trigger was Amazon researchers finding a method of bypassing Fable 5's safeguards by prompting it into identifying a number of software vulnerabilities. Anthropic's own testing then found that less capable models, including Claude Opus 4.8, GPT-5.5, and Kimi K2.7, could identify the same vulnerabilities, which is the interesting part: the jailbreak was not unique to Fable 5. The company says a new classifier blocks the specific technique described in the Amazon report in over 99% of cases. Pro, Max, Team, and select Enterprise plans will get Fable 5 for up to 50% of weekly usage limits through July 7.

The more strategic move is the joint work with Amazon, Microsoft, Google, and other Glasswing partners on what Anthropic calls a consensus framework for assessing the severity of AI jailbreaks. Four criteria are on the table: capability gain, breadth of capability gain, ease of weaponization, and discoverability. Alongside the rubric, Anthropic is launching a HackerOne program where security researchers can submit potential cyber jailbreaks they have discovered in Fable 5.

The honest caveat is that this is Anthropic's own post, so the specifics — the 99% figure for the classifier, the eventual shape of the severity rubric, the extent to which Microsoft and Google actually adopt it in their own products — are the company's claim rather than an independently verified state of the world. The post does not name a lead author for the framework, does not set a timeline for when the rubric becomes public, and does not explain how the participating labs plan to resolve disagreements when they score the same jailbreak differently.

What is interesting is the direction. A shared severity scale, if it holds together, would let researchers and buyers talk about jailbreaks the way they already talk about CVSS scores for regular software vulnerabilities. For enterprise buyers weighing model risk, that would be a real upgrade over the current signal, which is usually just the raw headline that a model got jailbroken.

What others are reporting

Coverage cluster as of 24h after publish

  1. Fortune Read →

    Frames the redeployment as a business-political truce, citing Anthropic's pending IPO as a pressure point that made a prolonged standoff with the Trump administration untenable.

    Our hope is that this collaboration, along with our proposed consensus industry framework, will serve as the basis for systematic rules for the whole industry.
  2. Forbes Read →

    Leads with the enterprise resilience angle: force majeure clauses did not cover government-mandated AI shutdowns, and model diversification is now a governance requirement.

    Model diversification is now a governance requirement for enterprises managing AI dependencies across cloud platforms.
  3. The Hacker News Read →

    Most detailed account of the four-axis jailbreak severity framework; notes the new classifier blocks the technique in 99%+ of cases but raises false positives on routine coding.

    The company calls the flagged behavior routine defensive security work, not a hidden super-capability.
  4. Eastern Herald Read →

    Sole outlet to center Katie Moussouris's expert dissent: the government's jailbreak characterization rested on verbal evidence and was never formally disputed before controls were lifted.

    That is not a guardrail bypass. It is the most valuable thing an AI model can do for defensive security.
  5. The Next Web Read →

    Notes the reversal's political opacity: the exact role of China competitive pressure remains unconfirmed, and the redeployment conditions read as the price Anthropic paid to get back online.

    Commerce lifted the controls, Lutnick set conditions, Anthropic accepted them, and access is returning.
  6. Crypto.news Read →

    Frames the episode as a policy clash: a single Amazon researcher report became the basis for a full Commerce Department shutdown that Anthropic contested throughout.

    We've received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.

Shared on Bluesky by 5 AI experts

Originally reported by anthropic.com

Read the original article →

Original headline: Anthropic Redeploys Fable 5 July 1 With July 7 Usage Credits — Also Unveils Joint Jailbreak Severity Standard With Amazon, Microsoft, and Google