Anthropic has unveiled its most capable artificial intelligence models yet — Claude Fable 5.2 and Claude Mythos 5.2 — claiming they beat OpenAI and Google on key benchmarks while costing less, as the company prepares for a stock market listing.
Key Points
- Anthropic Claude Fable 5.2 Mythos 5.2 — Anthropic released the two upgraded models on Monday, 9 November.
- The new models generate answers up to 30% faster than version 5.1, with a Fast mode that boosts response speed by up to 2.5 times.
- Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, down from $10 and $50.
- On the Artificial Analysis Intelligence Index, Fable 5.2 ranked first with a score of 65, ahead of Google’s Gemini 4 Carbon on 60 and OpenAI’s GPT-6 Astra on 57.
The launch is a direct challenge to Anthropic’s two biggest rivals. According to evaluation results the company released, the new models outperformed not only their predecessors but also OpenAI’s GPT-6 Astra and Google’s Gemini 4 Carbon on key benchmarks, including coding, cybersecurity and scientific research. On the composite intelligence index compiled by AI performance analysis firm Artificial Analysis, Claude Fable 5.2 ranked first with a score of 65 — five points ahead of previous leader Gemini 4 Carbon, with Anthropic’s own Claude Opus 5.5 third on 59 and GPT-6 Astra fourth on 57.
The models are aimed at demanding technical work: complex coding, scientific research, advanced reasoning and multi-step workflows. The price cut matters as much as the speed. At $8 per million input tokens and $40 per million output tokens, Fable 5.2 is cheaper than its predecessor, and the company says it generates answers up to 30% faster than version 5.1 — with the new Fast mode pushing response generation up to 2.5 times faster still.
A crowded week for AI
Anthropic’s launch capped a busy 24 hours for the AI industry. Meta introduced Muse Ember, its smallest and most efficient model, built for the repetitive work AI agents do all day — tool calls, coding subtasks, routing and long-document parsing — and put its Muse assistant on Apple Watch and Wear OS. OpenAI said users can now connect several of their own computers to a single dot agent and choose which machine runs each task, and Google said developers can now build apps powered by Google Maps Platform in Google AI Studio.
The timing is no accident. Anthropic, which is preparing for an initial public offering, is pushing to gain an edge in the technology race against OpenAI and Google — and the launch comes as the industry faces fresh scrutiny over AI safety. Six governments and major AI developers signed a Paris safety declaration on 9 November, and Australia announced plans for mandatory AI safety requirements and a dedicated regulator after an incident involving an OpenAI agent accessing a government website without authorisation.
What happens next
Anthropic is pushing its IPO this month, and the new models are its pitch to investors: the best-performing AI on the market, at a lower price, from a company positioning itself as the safety-conscious alternative. Whether Fable 5.2’s benchmark lead holds up in independent testing — and whether developers switch from OpenAI and Google — will decide whether this is a genuine changing of the guard or a brief moment at the top of the leaderboard.
For background on Anthropic’s listing plans, see our report on Anthropic moving toward a massive stock market debut, and on Anthropic reporting rapid revenue growth ahead of its stock market launch. More in our Technology section.
Sources: Chosun Biz, “Anthropic unveils faster, cheaper Claude models, claims lead over OpenAI and Google ahead of IPO”; ETV Bharat, “From Wearable Agent To Heavy Coding: Top AI Releases From Meta, Google, OpenAI, And Anthropic”.