AI models reportedly rival Anthropic’s Mythos at finding critical software flaws
- Bookmark
- CommentsGo to comments
A Chinese artificial intelligence company claims to have created a “cyber nuclear weapon” that could carry out devastating attacks on Western companies and governments.
Beijing-based Zhipu AI (Z.ai) said its AI model now matches Anthropic’s Mythos model at finding cyber security vulnerabilities, putting China on par with the US.
When Mythos was unveiled in April, Anthropic said that it had uncovered software vulnerabilities in “every major operating system and every major web browser” in the world, including some that had gone undetected for decades.
The US firm issued a limited roll out of the model in order to prevent bad actors from taking advantage of it to carry out cyber attacks. Earlier this month, Anthropic was forced to disable its Mythos 5 model for all users following a directive from the US Department of Commerce.
It has since been redeployed to around 100 trusted entities, which include corporations, critical infrastructure operators, and US government agencies.
According to Z.ai, its GLM-5.2 model can match the ability of Anthropic’s Mythos at finding software bugs, while operating at around one sixth of the cost.
Another Chinese cyber security firm, 360 Security Technology, also claims to have built AI models that can rival Anthropic’s Mythos.
Company founder Zhou Hongyi described Mythos as the AI era’s “cyber nuclear weapon”, and said that his company’s Yitian Tulong models were the “Chinese Mythos” at the ISC.AI 2026 security conference in Beijing last week.
The ideal summer spot? Away from scams.
Get All-in-One Protection for Your Digital Life
ADVERTISEMENT
The ideal summer spot? Away from scams.
Get All-in-One Protection for Your Digital Life
ADVERTISEMENT
“This kind of powerful weapon that can change the landscape of cyber offence and defence cannot be held only by others,” he said.
“China cannot wait until model capabilities have fully caught up before starting vulnerability discovery, because we cannot afford to wait.”
The capabilities of the models from the two companies are yet to be independently verified by Western firms.
Cyber security experts warned that the rhetoric could be meant as a deterrent against the US, rather than proof of such capabilities.
“It’s not surprising that China has publicly stated that they have a tool that is equivalent to Anthropic's Mythos, and linking this to the Nuclear Deterrent is quite an interesting tactic,” James Griffiths, who worked at UK intelligence agency GCHQ before founding cyber consultancy firm UtopianKnight, told The Independent.
“With the latest reporting coming out of the USA about it walking through NSA systems and able to breach even the highest classification networks clearly this is a power play by China using the guise of the ‘nuclear deterrent’ concept. Don't try to use this on us or we will use ours on you. Mutually assured self destruction in cyber space.”
Join our commenting forum #
Join thought-provoking conversations, follow other Independent readers and see their replies
Comments