Anthropic Unveils Claude Fable 5.2 and Mythos 5.2, 20% Cheaper
Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on Monday, November 9. The models run 30% faster and cost $8 per million input tokens. The launch comes as the company prepares an IPO that could happen as soon as next week.
Anthropic unveiled two AI models, Claude Fable 5.2 and Claude Mythos 5.2, on Monday, November 9, 2026, priced at $8 per million input tokens and $40 per million output tokens. Tokens are the small units of text that AI models read and write, and they are how providers charge for usage.
The company said the models generate output up to 30% faster than their 5.1 predecessors and communicate more clearly with users. The launch comes as Anthropic readies an initial public offering that could take place as soon as next week.
Who can use each model
Fable 5.2 is generally available. It targets coding, reasoning and AI agent tasks, meaning software that carries out multistep work on its own.
Mythos 5.2 is restricted. Anthropic said it is open only to organizations approved through its Life Sciences and Cyber Verification Programs, for cybersecurity and life-sciences research.
Anthropic first introduced Mythos in April and limited it to a select group of organizations because of its cybersecurity capabilities. Fable, a publicly available version with extra safeguards, followed two months later.
Price and performance claims
Anthropic said the new models cost 20% less than the 5.1 generation. Against OpenAI's GPT-6 Astra, which came out in early September, the new models are also cheaper: Astra costs $10 per million input tokens and $50 per million output tokens.
The gap is wider for cached tokens, which cover repeated material such as long documents. Anthropic charges 20 cents per million, versus $1 for Astra.
Anthropic said the models beat competitors on coding benchmarks, and that over 90% of its own code is written by AI. Those claims come from the company and have not been tested independently.
It also said a team of 1,000 agents powered by Fable 5.2 formalized the proof of Fermat's Last Theorem in 30 hours, against 11 days for an earlier unreleased research version of Claude. An unnamed security company, Anthropic said, found that one in six confirmed bugs found by Mythos 5.2 were missed by all other frontier models it tested.
OpenAI released a cybersecurity-focused model, GPT-6 Cyber, late last week.
Safety and the IPO
This is the first time Anthropic has released frontier models, its most advanced tier, since CEO Dario Amodei called in September for a slower pace of AI development. The company said the models went through rigorous safety and alignment testing, with new evaluations and post-launch monitoring tools.
"We called for pacing, not pausing," Dianne Penn, Anthropic's head of product management for research, said in an interview.
Penn said Fable 5.2 suits "power users or power use cases." For daily tasks, Anthropic points customers to Claude Opus 5.5, released in September. Fable 5.2 is meant for hard problems in engineering, mathematics and robotics.
Anthropic's amended prospectus shows $33.24 billion in revenue for the nine months ended September 30, up roughly 1,200% from a year earlier. The same filing revealed a third-quarter slowdown in revenue growth.
The company is aiming for a valuation of $1 trillion to $1.5 trillion in the offering, down from a $2 trillion target floated in recent months. Executives are preparing to begin the IPO roadshow, the series of investor meetings that precedes a listing.
