Anthropic's Fable 5 Rises
Anthropic (anthropic.com) announced on Tuesday that the U.S. Department of Commerce has lifted the export controls on its Claude Fable 5 and Mythos 5 models, ending one of the most aggressive regulatory interventions against a commercially deployed AI model to date. Fable 5 is now available again to global users on the Claude platform, Claude.AI, and Claude Code.
Anthropic launched Claude Fable 5 on June 9, 2026, as the first publicly available model in its Mythos-class tier, which exceeds the capability of the Opus line. The model is state-of-the-art on nearly all tested benchmarks, with exceptional performance in software engineering, knowledge work, vision, and scientific research. Three days later, the ride was over. Commerce Secretary Howard Lutnick sent Anthropic CEO Dario Amodei a letter directing the company to restrict both models under export controls, citing national security authorities. The directive required Anthropic to suspend access by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. Because the company could not verify users' nationality in real time, it took both models offline for everyone.
The catalyst was a jailbreak. A trusted partner, which CNN reported was Amazon, found a way to get around Fable 5's guardrails. Mythos 5 is the powerful cybersecurity model, and Fable 5 is the consumer-facing layer built on top of it, so a bypass of one raised alarms about the other. Anthropic pushed back on the severity, saying at the time that the jailbreaks were "simple" and that other publicly available models had similar workarounds. But the government initially held firm, and the standoff dragged on for nearly three weeks.
The resolution came in stages. Last Friday, Lutnick gave Anthropic permission to release Mythos 5 to a select group of companies and federal agencies. As CNBC reported, a letter viewed by the outlet said the Commerce Secretary had determined that appropriate safeguards were in place to permit certain trusted partners to access the model. Then on Tuesday, Lutnick confirmed the full reversal.
"Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America's leadership in AI."
As a condition for restoring access, Anthropic agreed to proactively detect and address security risks, help develop standards for future models, and report malicious activity to the government. The company also said it trained a new safety filter that blocks the reported workaround more than 99% of the time. Anthropic is continuing to work with the government to expand access to Mythos 5 for additional domestic and international partners through Glasswing, its cybersecurity initiative that provides selected organizations with advanced AI models for defensive security testing.
The episode leaves open the question of what framework should govern future frontier model releases. President Trump signed an AI executive order in early June asking developers to voluntarily submit models for government capability assessments before a full release, but federal agencies have 60 days to establish the relevant processes. Meanwhile, Chinese open-weight models are gaining ground fast, and they ship weights without full visibility into training data or methodology, raising the risk of embedded backdoors or unexpected behaviors in models that are being widely adopted.
Anthropic raised $65 billion at a $965 billion valuation just weeks before the Fable 5 launch, and is also separately fighting the Pentagon over its designation as a "supply-chain risk." The company has been at loggerheads with the administration for much of the year.
The broader industry is also feeling the effects. OpenAI announced Friday that it is launching its latest model series, including GPT-5.6, for a small group of trusted partners first, previewing the models' capabilities with the government ahead of release. OpenAI said it does not believe this kind of government access process should become the long-term default, noting that it keeps the best tools from users, developers, and cyber defenders who need them.
In its brief window before the ban, Fable 5 had already generated significant excitement among developers and researchers who found it to be a step change in coding, scientific reasoning, and knowledge work. Its return is likely to be one of the most closely watched model launches of the year.