Anthropic launches Fable 5.2 and Mythos 5.2 AI models ahead of IPO

Anthropic PBC released new versions of its Fable and Mythos artificial intelligence models on Monday as the company prepares for its initial public offering later this month.

Fable 5.2 can perform more complex software engineering tasks, control robots and manage long-running projects, the company said. The model can coordinate over 1,000 AI agents simultaneously. Mythos 5.2, which has fewer safety restrictions for cybersecurity and life sciences applications, will be available to select organizations through a verification program.

Both models operate faster than previous versions, and Fable 5.2 costs less than its predecessor, Anthropic said.

The company first introduced Mythos in April, limiting access to government agencies, financial institutions and major software providers due to the model's ability to identify and exploit security vulnerabilities in critical software. Fable launched two months later, offering similar capabilities to a broader market while blocking cybersecurity tasks.

The US government restricted access to both models in June over safety concerns. Access to Fable 5 was restored after Anthropic added classifiers to detect and block cybersecurity tasks. Mythos 5 access was later reinstated for a limited group of users.

Dianne Penn, who leads research product management at Anthropic, said the company extended testing periods for the latest models and used additional evaluation methods. "Pacing is not like pausing," Penn said.

Anthropic is expanding access to Mythos, making the new version available to companies in the US and internationally, Penn said.

Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, down from $10 and $50 for Fable 5.1.

In mid-September, Anthropic Chief Executive Officer Dario Amodei urged top AI firms to reduce the pace of developing their most advanced models.