Anthropic’s Claude completes first machine-verified Lean proof of Fermat’s Last Theorem in 11 days
Anthropic announced that its AI model Claude autonomously produced the first fully computer-verified proof of Fermat’s Last Theorem in 11 days. The formalization involved writing 13 million lines of Lean code and proving 30,300 theorems, ultimately using 29,500 of them. This is over five times the scale of Mathlib, the standard Lean mathematical library, and marks a major milestone in AI-driven mathematical research.
Editorial responsibility
- No named human review is recorded for this page.
- Reports are grouped by semantic similarity and deterministic rules. Language models may assist titles, summaries, translation and cross-source analysis; the page itself is projected from evidence records.
- Current automated evidence projection
Cross-source coverage
Wire timeline
Anthropic: Claude completes first fully computer-checked proof of Fermat's Last Theorem in 11 days
Anthropic announced that its AI model, Claude, completed the first fully computer-checked proof of Fermat's Last Theorem in 11 days. The achievement involved translating Andrew Wiles' 1995 proof into a form verifiable by computer, a task far more complex than simple code translation. Working largely autonomously, dozens of Claude agents produced 13 million lines of Lean code and roughly 29,500 supporting theorems across algebra, geometry, number theory, and harmonic analysis. This includes mathematics that had never been formalized before. The development could help mathematicians check new research faster, catch hidden gaps, and rigorously evaluate AI-generated proofs.
Claude compresses years-long Fermat's Last Theorem formalization into 11 days
A post on X by rohanpaul_ai reports that AI, specifically Claude, has dramatically accelerated the formal verification of mathematical proofs. The post explains that normal mathematical proofs leave thousands of steps unstated, which human mathematicians can infer but formal systems like Lean cannot. Converting a proof like Fermat's Last Theorem into machine-checkable pieces was expected to take years of human work. However, Claude compressed that years-scale formalization task into just 11 days. The post argues this breakthrough will make formal verification practical for much more mathematics, addressing a critical bottleneck as AI produces far more proofs than humans can manually review.
Claude formalizes Fermat's Last Theorem in 11 days, producing 13 million lines of Lean code
An AI system, Claude, has formalized Fermat's Last Theorem in 11 days, a task expected to take years. The process involved dozens of Claude agents converting the existing Wiles-based proof into Lean code, producing 13 million lines of Lean and 29,500 intermediate theorems. The Lean proof assistant successfully verified the finished proof. This achievement demonstrates AI's potential to automate the labor-intensive work of turning advanced human mathematics into machine-checkable proofs, marking a significant advance in automated mathematical reasoning.
Show 2 older updatesHide older updates
Claude completes first formalized proof of Fermat's Last Theorem, totaling 13 million lines
AnthropicAI announced that its AI model, Claude, completed the first formalized proof of Fermat's Last Theorem, one of the most famous theorems in mathematics. The proof, totaling over 13 million lines of code, provides machine verification of the theorem first proven by Sir Andrew Wiles in 1995. The project, which experts thought would take many years, also proves over 29,000 other theorems across many areas of mathematics that had never before been formalized. This represents the largest Lean proof ever written and is seen as a major step in firming up the core of mathematical knowledge. AnthropicAI expressed optimism that AI-assisted verification of mathematical proofs will help reduce the burden of refereeing mathematics in an era of increasing proof production. The complete proof is available on GitHub, and a detailed process description is available on AnthropicAI's Science Blog.
Anthropic's Claude autonomously produces first machine-verified Lean proof of Fermat's Last Theorem
Anthropic announced the first fully computer-verified proof of Fermat's Last Theorem, completed largely autonomously by its AI model Claude in 11 days. The formalization involved writing 13 million lines of Lean code and proving 30,300 theorems, ultimately using 29,500 of them. The scale of this effort is over five times that of Mathlib, the standard mathematical library for Lean. The original publication details the formalization path, the Prove2Me platform mechanism, and code scale, illustrating a feasible approach for AI to automatically formalize large-scale mathematical proofs. This achievement marks a significant milestone in AI-driven mathematical research and automated theorem proving.