Hours after Anthropic claims a Claude proof of the Riemann hypothesis, Stanford releases an all-human competing draft

What happened
On Friday, Levent Alpöge, a mathematician at Anthropic, announced that he used the Claude AI model to produce a proof of the Riemann Hypothesis, also extending the result to the Generalized Riemann Hypothesis. He released a paper and computer code documenting the work. Within hours, Stanford mathematicians, led by professor Ravi Vakil and mathematician Alessandro Masullo, published a competing draft proof created entirely by humans. The Stanford team released their draft earlier than planned in response to Anthropic's announcement.
Key details
- Levent Alpöge worked with the Claude AI model over several months to develop the proof.
- Alpöge's AI-generated proof, if printed, would exceed 250,000 pages and was released in the Lean programming language, split into 14 data packages.
- Stanford's Alessandro Masullo had his draft reviewed for months by number theorist Brian Conrad before its rushed release.
- Several mathematicians and AI models quickly identified issues in the Stanford draft, which are now under review.
- The Riemann Hypothesis is one of the Clay Mathematics Institute's Millennium Prize Problems, carrying a $1 million reward.
Why it matters
Professional mathematicians, research institutions, and the AI community are directly affected, as two separate groups have now released competing proofs of a major unsolved problem. The Anthropic proof relies on AI-generated reasoning and formal computer verification, while the Stanford draft represents a traditional, human-only approach. Both proofs require rapid review and validation, creating immediate work for experts and potentially changing how mathematical breakthroughs are achieved and verified.
