Anthropic has introduced two of its most powerful AI models yet, Claude Fable 5.2 and Mythos 5.2, just over two months after launching their 5.1 versions. According to the company, the new models are considerably better at coding, reasoning, and scientific research. They can also generate responses up to 30% faster than their predecessors, while costing less to run. Anthropic has also shared some pretty impressive examples of what the new models have managed to accomplish.
Fable 5.2 posts massive improvements across key benchmarks
Based on benchmark figures released by Anthropic, Fable 5.2 scored 49.1% on FrontierCode 2.0, a big jump from Fable 5.1’s 32.4%. It also raced ahead on scientific research tasks, scoring 74.6% against its predecessor’s 52.6%. The biggest improvement, however, came in robotics simulations, where its score jumped from 27.6% to 64.9%.
Anthropic has also demonstrated some rather ambitious applications of its new model. One of its engineers tasked Fable 5.2 with developing an operating system from scratch. Over eight days, it coordinated more than 1,000 AI agents and produced a working operating system that can boot on a laptop.
The company also claims Fable 5.2 solved two previously open problems related to black holes and helped formalize the mathematical proof of Fermat’s Last Theorem in just 30 hours.
Mythos 5.2 is more capable, but access remains restricted
Just like the previous generation, Fable 5.2 and Mythos 5.2 share the same underlying model. However, Mythos has fewer safeguards for sensitive cybersecurity and biomedical research. Anthropic claims Mythos 5.2 is significantly better at finding software vulnerabilities, scoring 89.6% on its CVE Rediscovery benchmark against Mythos 5.1’s 78.3%.
Anthropic is also cutting the price of both models to $8 per million input tokens and $40 per million output tokens, down from $10 and $50. There’s also an optional fast mode that generates tokens up to 2.5 times faster, although it costs twice as much.
Fable 5.2 is available to regular Claude users and developers, but Anthropic is keeping Mythos 5.2 on a short leash. Access to the latter requires approval through its cybersecurity or life sciences verification programs, although Enterprise customers can also use its capabilities through Claude Security.