Anthropic on Monday released Claude Fable 5.2, a new top-tier AI model available to the public, along with Claude Mythos 5.2, which is limited to vetted organizations at launch. The company says both models are faster and cheaper than their 5.1 predecessors, according to The Verge, 9to5Mac and CNBC.
Anthropic says the models generate output up to 30% faster than the 5.1 versions. CNBC reported that they cost 20% less. Pricing is $8 per million input tokens and $40 per million output tokens.
Anthropic describes the models as built for “daunting” technical challenges. It claims improvements in coding, reasoning and vision, and says testers found they communicate more clearly than earlier versions.
The company said that during prerelease testing, the models helped engineers complete software migrations that would otherwise have been infeasible. It also said they designed three working electronic devices, solved two open problems in black hole physics and formalized the proof of Fermat's Last Theorem in 30 hours. These are Anthropic's own claims, as reported by The Verge.
On coding, Anthropic said more than 90% of the changes merged into its core codebase are now written by Claude with human guidance. It said a team of agents running an earlier version of Fable 5.2 migrated a critical production service from Python to Rust. According to the company, the service now starts about 26 times faster.
Dianne Penn, Anthropic's head of product management for research, told CNBC that Fable 5.2 is aimed mainly at experienced users working on complex, open-ended tasks in engineering, mathematics and robotics. She also said the company had called for managing the pace of AI development rather than a complete pause, according to CNBC.
Anyone on a plan that includes Fable models can use Fable 5.2 now. Mythos 5.2 is not open to general users at launch.
Developers paying per token face the new rates of $8 per million input tokens and $40 per million output tokens. Anthropic is also giving subscribers on eligible paid plans extra Fable usage this week, according to 9to5Mac. Plans and seats that include Fable get one to three Fable resets, depending on how long the customer has subscribed. Fable is included for up to 50% of weekly usage limits on Pro plans and standard Team seats.
Anthropic does not recommend Fable or Mythos as the default for everyone. For most coding and knowledge work, it suggests starting with Claude Opus 5.5 and moving up to Fable or Mythos, or down to Claude Sonnet 5.5 or Claude Haiku 5.5, depending on the task.
CNBC reported that the launch comes as Anthropic prepares for an initial public offering that could happen as early as next week. The company is seeking a valuation of between $1 trillion and $1.5 trillion. In an amended prospectus, Anthropic said revenue for the nine months ended Sept. 30 was $33.24 billion, about 1,200% more than a year earlier.
Anthropic said the new models went through extensive safety and alignment testing before release. It also said it developed new evaluation methods and post-launch monitoring tools, according to CNBC.
Read next
Is DoorDash Buying Instacart? What the Reported Exclusive Talks Show
What a Next-Gen Xbox Priced Up to $1,800 Could Mean for Console Buyers
GTA 6 Achievement Leak Claims 55 Entries, One Worth 1,000
Around Prague, based on reporting by The Verge, 9to5Mac, UA.NEWS