Asian AI startups launch Mythos-like models as Anthropic’s export ban drags on
TL;DR
Tokyo-based Sakana AI launched Fugu, an orchestration model for agents that it says can stand with Anthropic Fable 5 and Mythos Preview. China’s 360 introduced Tulongfeng and Yitianzhen for vulnerability discovery, cyber defense, and incident response. Its framing is more openly strategic than Sakana’s. The U.S. export restriction on Anthropic Mythos and Fable has been in place for about two weeks, creating a market gap that Asian providers are now targeting.
Nauti's Take
Export controls are meant to preserve leverage, but they can also redirect demand. That is the live experiment here: the market is learning that a slightly weaker but reliably accessible model can be strategically better than the best model behind a political gate.
For Anthropic and other U. S.
labs, the risk is that infrastructure customers do not simply return when a ban ends. Once workflows, data paths, and procurement move local, new dependencies harden fast.
Briefingshow
The key issue is not only whether Fugu or Tulongfeng truly match Mythos-class capability. Availability itself is becoming a product feature. If companies or governments believe a U.
S. model can disappear overnight, local alternatives gain trust, data access, and integration depth much faster.
Sources