Anthropic, the artificial intelligence company behind Claude, has begun testing a new AI model more capable than any previously released, Fortune reported.
The company said the model represents “a step change” in performance and is “the most capable we’ve built to date.” It is currently being tested with a small group of early access customers while Anthropic evaluates its behavior and risks.
Among the names that fell sharply on the news: Palo Alto Networks (PANW), Crowdstrike (CRWD), and Fortinet (FTNT) fell between 4% and 6%. The broader iShares Expanded Tech-Software Sector ETF (IGV) is down 2.5%.
The overnight leak likely contributed to bitcoin’s decline. falling back to $66,000 after flirting with $70,000 hours earlier.
Details about the model emerged after internal materials were accidentally exposed in a publicly accessible data warehouse, according to Fortune. About 3,000 assets linked to Anthropic’s blog were available online, including draft ads and internal content that had not yet been published.
Among the files was a draft blog post that referred to the model as “Claude Mythos.” The document warned that the system could pose serious cybersecurity risks, noting its ability to identify and exploit software vulnerabilities.
Anthropic currently offers three levels of models (Opus, Sonnet and Haiku) that vary in size, cost and capacity. Leaked materials suggest that the company is developing a new level called “Capybara,” which would be even bigger and smarter than Opus, the company’s most advanced model to date.
