By Interestana AI Editorial — AI-drafted, human-overseen. How we report
Anthropic AI Formalizes Fermat's Last Theorem Proof
Anthropic's artificial intelligence model, Claude, has formalized a computer-checked proof for Fermat's Last Theorem, a significant mathematical conjecture that remained unproven for centuries. This achievement was published online in Nature on September 7, 2026, with the digital object identifier 10.1038/d41586-026-02822-9. The AI model generated a proof consisting of 13 million lines of code, which was then verified by a computer. This complex task was accomplished by Claude in an remarkably short timeframe of just 11 days.
Fermat's Last Theorem, first stated by Pierre de Fermat around 1637, posits that no three positive integers a, b, and c can satisfy the equation a^n + b^n = c^n for any integer value of n greater than 2. Despite Fermat's claim to have a "truly marvelous proof" which he could not fit into the margin of a book, the theorem remained unproven for over 350 years. The first complete proof was finally established by mathematician Andrew Wiles in 1994, utilizing advanced concepts from algebraic geometry and number theory. The formalization by Anthropic's AI represents a new method of tackling such profound mathematical problems, moving beyond human intuition and manual verification to automated, rigorous checking.
The development highlights the growing capabilities of advanced AI models in abstract reasoning and complex problem-solving, areas traditionally considered exclusive to human intellect. Claude's ability to construct and verify such an extensive and intricate proof signifies a potential paradigm shift in mathematical research. This accomplishment could pave the way for AI to assist mathematicians in exploring new theorems, verifying existing proofs with unprecedented speed and accuracy, and potentially uncovering novel mathematical insights that might elude human researchers. The 13-million-line proof generated by Claude is a testament to the model's sophisticated understanding and manipulation of mathematical logic.
This milestone is particularly noteworthy given the abstract and theoretical nature of Fermat's Last Theorem. The theorem's complexity requires a deep understanding of advanced mathematical principles, and its proof involves intricate logical deductions. For an AI to not only comprehend these principles but also to construct a verifiable proof within a matter of days is a remarkable demonstration of its computational and reasoning power. The collaboration between Anthropic and the mathematical community, as evidenced by the publication in Nature, underscores the potential for AI to become an indispensable tool in scientific discovery and advancement, pushing the boundaries of what is currently understood and achievable in fields like mathematics.
Original source — read the full reporting at the publisher:
Read on NatureGet the weekly AI digest
AI news + new model releases, weekly. Drafted by our agents, reviewed by humans.