Anthropic CEO Dario Amodei warned there is little time left to fix tens of thousands of software vulnerabilities found by the company's AI model Mythos, CNBC reported on Monday.

Amodei made the remarks at an Anthropic-hosted event where he appeared on stage with JPMorgan Chase CEO Jamie Dimon.

He said Chinese AI models lag Anthropic's products by 6 to 12 months, and that period is the time available to fix the vulnerabilities.

Mythos, which was made available on a limited basis to some organisations last month, drew attention by finding major software vulnerabilities that had gone undiscovered for decades.

Amodei said an earlier Anthropic model found about 20 vulnerabilities in the Firefox browser, but Mythos found nearly 300. He said the total number of vulnerabilities it found across all software runs into the tens of thousands.

Most of the vulnerabilities found by Mythos are not publicly known, and no patches have been made yet.

Amodei warned that if the vulnerabilities are disclosed, "malicious attackers will exploit them." Anthropic is restricting access to Mythos to a small number of partner companies due to concerns it could be abused by criminals or hostile states. Amodei warned that "there is a risk that the number of vulnerabilities and breaches will surge and that financial damage from ransomware will rise sharply, hitting not only schools and hospitals but also banks."

Anthropic also announced an expansion of its financial services platform on Monday. It introduced 10 new AI agents that automate investment banking and back-office work, and added integration with Microsoft Office programs.