Dario Amodei, co-founder and CEO of Anthropic, during an interview at Anthropic headquarters in San Francisco, California, USA, on Thursday, April 30, 2026.
Jason Henry | Bloomberg | Getty Images
human The company on Monday released two new artificial intelligence models, Claude Fable 5.2 and Claude Mythos 5.2, as it prepares for a historic IPO that could take place as early as next week.
Fable 5.2 and Mythos 5.2 are Anthropic’s most advanced products, and the company says they produce output 30% faster than previous versions. The company also said that prices for these models will be 20% lower than the 5.1 generation, and will be available at $8 per million input tokens and $40 per million output tokens.
The announcement marks the first time Anthropic has released a so-called “Frontier” model since CEO Dario Amodei called for a slowdown in AI development in September. Anthropic said the model has undergone rigorous safety and conditioning testing, and the company has developed new evaluation tools and post-launch monitoring tools ahead of launch.
“We asked for a pace, not a pause,” Diane Penn, head of research product management at Anthropic, told CNBC in an interview.
Anthropic must balance its commitment to safety with intense commercial and competitive pressures, a tension the company highlighted late last month in the broader risk factors section of its prospectus. And as management prepares to launch Anthropic’s IPO roadshow, it will need to convince investors that its approach is the right one.
As CNBC previously reported, the company is targeting a stock valuation of $1 trillion to $1.5 trillion, below the $2 trillion target floated in recent months.
Anthropic first introduced Mythos in April, but the model’s advanced cybersecurity features limited deployment to a select group of organizations. Two months later, the company announced Fable, a generally available version of Mythos with additional protections.
These models, particularly the Mytos, have captivated Wall Street and government officials alike and helped fuel Anthropic’s rapid growth this year. The company said in its revised prospectus that it had revenue of $33.24 billion for the nine months ended Sept. 30, an increase of about 1,200% from the same period last year.
Peng said Fable 5.2 is primarily aimed at “power users or power use cases” as it excels at complex and free technical work. Other models, such as the Claude Opus 5.5 that Anthropic released in September, are better suited for day-to-day tasks, and users should use the Fable 5.2 when they need to tackle difficult problems in engineering, math, robotics, etc., Penn said.
“These are completely new and frontier use cases,” Penn said. “Over time, we expect it to probably trickle down to the Opus class, but right now we’re seeing a difference in terms of performance, and we think these frontier use cases will benefit.”
Anthropic said Fable 5.2 will be generally available on Monday, and Mythos 5.2 will be made available to vetted organizations through the company’s life sciences and cyber validation programs.
WATCH: Anthropic’s incredible rise: What you need to know before the IPO
