Mythos 5 is partially back, but Fable 5 is still gone. And the two-week gap gave Asian competitors enough time and narrative to establish themselves as export-control-safe alternatives. The window may close, but the market signal will not.
- The US government has authorized Anthropic to restore Mythos 5 access for more than 100 US companies and government agencies, including their non-American employees.
- Fable 5 was not addressed in the directive and remains off the market globally. Anthropic says it is working with the government to restore general access.
- While the ban ran for two weeks, Japan’s Sakana AI and China’s 360 both launched frontier cybersecurity AI models positioned as export-control-free Mythos alternatives.
- 360’s Tulongfeng is designed to automatically discover software vulnerabilities. Its companion model Yitianzhen automates cyber defence and incident response.
- Sakana AI positioned Fugu as a hedge, not a replacement, saying US models remain important to Asia. 360 framed it as a national strategic asset.
What Happened
The US government has authorized Anthropic to restore access to Mythos 5 for more than 100 US companies and government agencies, two weeks after the export ban forced Anthropic to pull both Mythos 5 and Fable 5 from the market, TechCrunch reported.
Commerce Secretary Howard Lutnick wrote to Anthropic’s chief compute officer Tom Brown authorizing the restoration, including access for non-American employees at those organizations and Anthropic’s own non-American staff. Fable 5 was not addressed in the directive and remains unavailable globally. Anthropic confirmed the development on X, saying it is continuing to work with the government to expand Mythos 5 access and restore Fable 5 for general use.
While the ban played out, Asian AI companies moved quickly. Japan’s Sakana AI launched Fugu, its multi-agent orchestration model, marketing it as a frontier-level alternative without export control risk. Sakana’s website advertises “delivering frontier capability without the risk of export controls.”
China’s 360 unveiled two AI security tools: Tulongfeng, designed to automatically discover software vulnerabilities, and Yitianzhen, built to automate cyber defence and incident response. Both claim Mythos-level capabilities. 360’s founder framed vulnerability-finding AI as a national strategic asset and flagged the risk of “one-way transparency” in AI access, TechCrunch reported.
Why It Matters
The 15-day ban created a window that competitors actively exploited and branded around. Sakana AI built its marketing positioning around export control immunity. 360 framed AI cybersecurity as a national sovereignty issue. Even if the Mythos and Fable 5 ban fully resolves, the credibility those labs established with Asian enterprise and government customers during the gap will not automatically disappear. US AI labs operating in cybersecurity and regulated sectors now have a market education problem they did not have two weeks ago.
Anthropic’s partial restoration is not a full return to normal. Fable 5 remains unavailable globally, and access to Mythos 5 is still restricted to a vetted list. Sakana AI’s own spokesperson said US models remain important to Asia, suggesting there is no permanent realignment yet. But even a temporary gap is enough to accelerate local alternatives that were already in development and give buyers a reason to diversify. Relve, an AI trends intelligence platform, is tracking this as the first concrete test of how US export restrictions reshape global AI vendor relationships.
Since June 12, we’ve been working closely with the US government to restore access to Claude Mythos 5 and Fable 5. Today, the government notified us that Mythos 5, our strongest cybersecurity model, can be redeployed to a set of US organizations that operate and defend critical…
— Anthropic (@AnthropicAI) June 27, 2026
“Access to top models can disappear overnight.”
David Ha, Co-founder and CEO, Sakana AI
Bottom Line
Watch whether Fable 5 is restored in the coming weeks and how broadly the Mythos 5 authorized list expands. The pace of access restoration will tell you whether the government views this as a temporary security review or the beginning of a permanent licensing regime for frontier cybersecurity AI.
For SaaS founders and security engineering leads evaluating AI tooling, the two-week ban made one thing clear: single-vendor dependency on export-controlled models is a real operational risk. The Asian alternatives that emerged are not yet proven in enterprise production, but the vendor diversification conversation is no longer optional for any team with global operations.
