Published on
Anthropic is working on a new powerful artificial intelligence (AI) model that “poses unprecedented cybersecurity risks,” according to a leak from the company.
Last week, a data leak revealed that Anthropic was working on a new, powerful model, which the AI company has since acknowledged is called “Claude Mythos”.
An Anthropic spokesperson told Fortune it marks a “step change” in AI and it is the “most capable” model it has built so far with “meaningful advances in reasoning, coding, and cybersecurity”.
The technology could be so strong that it could be a hacker’s dream device. Cybersecurity stocks slumped following the Anthropic rumours.
Meanwhile, Anthropic is privately warning top government officials that Mythos makes large-scale cyberattacks much more likely in 2026, Axios reported.
Here is everything we know about Mythos.
How was the leak revealed?
A configuration error, or human error, in Anthropic’s content management system accidentally exposed a draft blog that described the model.
The blog, cited by Fortune, which was “available in an unsecured and publicly-searchable data store,” said that the new model “poses unprecedented cybersecurity risks”.
What can Mythos do?
Mythos is said to be a part of Capybara, which is a new tier of AI models t
Fortune cited Antropic as saying in the document, “‘Capybara’ is a new name for a new tier of model: larger and more intelligent than our Opus models—which were, until now, our most powerful.”
The model’s advanced capabilities appear to suggest the AI systems can tackle more intricate problems with more autonomy and precision.
Capybara and Mythos appear to refer to the same underlying model.
How will it affect cybersecurity?
As AI companies release improved models, AI agents- which are learning to act and reason without human input- get better too.
Hackers can thereby run multiple hacking campaigns at once, which becomes more difficult to protect against.
At the same time, employees are using AI agents which can unknowingly connect to their own work systems, which gives cybercriminals a door to enter.
Additionally, human identities are now easier to breach thanks to AI, a Palo Alto Networks executive told Euronews Next in March.
When will we see Mythos?
Mythos is extremely compute-intensive and expensive to run. Anthropic said it is working on making it much more efficient before any general release.
Euronews Next has reached out to Anthropic for comment.