Chinese and Japanese companies launch AI models, claim they are as powerful as Anthropic’s Mythos

Two AI companies in Japan and China have launched AI models for cybersecurity and claim that their new software matches the power of Anthropic's restricted “Mythos” and “Fable 5” models even as the US government maintains its strict export ban on Anthropic's most advanced piece of technology.

The sudden regional product launches come just two weeks after the Trump administration deployed emergency national security export controls, abruptly pulling Anthropic’s cutting-edge systems offline for all non-American users.

Japan’s Sakana AI rolls out ‘Fugu’

In Tokyo, prominent AI startup Sakana AI launched a new frontier model called “Fugu” – named after the Japanese word for blowfish. The company claims that Fugu “stands shoulder-to-shoulder” with Anthropic’s banned models, but with one critical geopolitical advantage: it delivers top-tier capabilities entirely free from the risk of American export controls.Co-founded in 2023 by former Google researchers Ren Ito, Llion Jones and David Ha, Sakana specialises in building efficient generative AI models tailored for Asian languages and cultures, according to a report by TechCrunch .Fugu is engineered to act as an “orchestration model” for autonomous AI agents. Rather than just answering prompts, Fugu can automatically coordinate and manage tasks across multiple other AI models using their APIs.

“Access to top models can disappear overnight. Relying on a single provider for national infrastructure is a risk... Collective intelligence is the practical hedge against this concentration of power,” Sakana CEO David Ha wrote on X. A spokesperson for Sakana told TechCrunch that the timing of the launch was “entirely coincidental” based on research planned since last year.

China’s 360 launches Cyber AI model and tool

While Japan's Sakana is framing its model as a defensive backup plan for US allies, Chinese cybersecurity giant 360 took a much more aggressive stance.

The Beijing-based firm unveiled two powerful new AI security systems: one is Tulongfeng an AI model engineered to go head-to-head with Anthropic's Mythos by automatically scanning and discovering hidden software vulnerabilities.And there is Yitianzhen, which is a tool built to automate rapid cyber defense and real-time incident response. According to news agency Reuters, 360 founder Zhou Hongyi described vulnerability-detecting AI as a “national strategic asset.”

$47 billion market in tough spot

The rapid emergence of local alternatives may dent American dominance in the region. Prior to the government intervention, Anthropic was on a historic financial trajectory, reporting that its run-rate revenue crossed a staggering $47 billion in May 2026.