An undisclosed Claude research model has achieved a significant breakthrough in tackling the Riemann Hypothesis, pushing the lower bound for the proportion of its non-trivial zeros on the critical line from 41.6% to 67.2%. This advancement, accomplished with the assistance of approximately 60 AI agents and 31 million tokens, represents a substantial leap over previous human-driven progress on this long-standing mathematical challenge. While not a complete proof of the hypothesis, the model independently explored research avenues, synthesized existing knowledge, and even generated a Lean version of its findings, showcasing advanced AI capabilities in open-ended problem-solving. AI
IMPACT Demonstrates AI's growing capability in open-ended scientific discovery, potentially accelerating progress in complex fields beyond standard benchmarks.
RANK_REASON The item details a research milestone achieved by an AI model on a complex mathematical problem, including details about the methodology and potential future implications. [lever_c_demoted from research: ic=1 ai=1.0]
- Anthropic
- Brian Conrey
- Claude
- Claude Code
- Eric Easley
- Fable 5.1
- GPT-6
- Jarred
- Levent Alpöge
- Riemann Hypothesis
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
