Anthropic says during prerelease testing Mythos 5.2 and Fable 5.2 formalized the proof of Fermat's Last Theorem and solved two open black hole physics problems

Formal mathematical proofs and complex physics problems can now be solved by AI, which will accelerate scientific discovery.

Anthropic has launched Claude Fable 5.2, a new AI model designed for complex technical and scientific tasks. A second model, Claude Mythos 5.2, is also available but restricted to vetted organizations. The company claims these 5.2 models demonstrate significant improvements in coding, reasoning, and speed compared to their 5.1 predecessors. During prerelease testing, these models reportedly assisted engineers with complex software migrations, designed three electronic devices from the ground up, and made progress on robotics evaluations. Notably, Anthropic states that Fable 5.2 and Mythos 5.2 formalized the proof of Fermat's Last Theorem within 30 hours and solved two previously open problems in black hole physics. Fable 5.2 is available immediately, while Mythos 5.2's access will be broader over time.

Anthropic's announcement of Claude Fable 5.2 and Mythos 5.2 signifies a leap in AI's capability to handle abstract reasoning and formal logic, moving beyond pattern recognition to genuine problem-solving in theoretical domains. The models' ability to formalize Fermat's Last Theorem proof and tackle open problems in black hole physics suggests a new era of AI-assisted scientific research, potentially reducing the time and human effort required for breakthroughs. This advancement could democratize high-level scientific inquiry, allowing researchers with fewer specialized resources to contribute to complex fields.

The implications for software engineering are equally profound, with claims of orchestrating agent teams and executing long-running projects, alongside designing electronic devices from scratch. This indicates a shift towards AI agents capable of not just writing code but managing complex development workflows and conceptualizing hardware. For organizations seeking to leverage advanced AI, the tiered release strategy for Mythos 5.2 suggests Anthropic is carefully managing the rollout of its most potent capabilities, likely anticipating significant demand and potential ethical considerations for such powerful problem-solving tools.