Prime Highlights
- China’s 360 and Tokyo-based Sakana AI have both launched models they claim match Anthropic’s restricted Mythos and Fable 5 tools.
- Sakana AI is positioning its Fugu model as a hedge against AI access disruption rather than a permanent replacement for US models.
Key Facts
- Sakana AI was co-founded in 2023 by former Google alumni and builds affordable generative AI models optimised for Japanese language and culture.
- China’s 360 is a cybersecurity firm whose founder framed advanced vulnerability-detection AI as a national strategic asset.
Background
Two Asian AI companies have launched products they claim rival Anthropic’s most advanced models, as a US government export ban on the American AI lab’s top tools enters its third week.
Chinese cybersecurity firm 360 unveiled two AI security tools in the fourth week of June. Tulongfeng is designed to automatically discover software vulnerabilities, while Yitianzhen is built to automate cyber defence and incident response. The firm’s founder described vulnerability-finding AI as a national strategic asset and warned against what he called “one-way transparency,” where only some actors can access advanced vulnerability-detection capabilities.
Tokyo-based Sakana AI launched Fugu, a frontier model named after the Japanese word for blowfish, around the same time. The company says Fugu stands alongside leading models, including Anthropic’s Fable 5 and Mythos Preview. Designed for agent orchestration, Fugu can coordinate access to other models through their APIs.
Sakana, co-founded by former Google alumni David Ha and Llion Jones alongside former Mercari and Stability AI executive Ren Ito, targets Japanese businesses and government agencies seeking to reduce exposure to tightening export controls. Its website advertises frontier capability without the risk of export restrictions.
A Sakana spokesperson said the timing of Fugu’s launch was coincidental, though the company acknowledged the moment brought more attention than expected. Co-founder Ren Ito has urged the US government to prioritise preserving AI access for its closest allies rather than restricting it.
Anthropic’s run-rate revenue crossed $47 billion in the second month of 2026’s second quarter.