DeepMind's AlphaProof Nexus proof agent paper published in Science, already proving research-level math
pushmeet · x · 2026-10-09
Pushmeet Kohli of Google DeepMind announces that the AlphaProof Nexus technical paper has been published in Science. Highlights:
- Predecessor AlphaProof achieved silver-medal performance at IMO 2024.
- AlphaProof Nexus is an LLM-powered proof agent; collaborator Gergely Bérczi of Aarhus University has used it to prove research-level results.
- The paper was led by Swarat Chaudhuri and colleagues.
- DeepMind credits the formal math community (Kevin Buzzard, Lean's Mathlib) for making AI proofs trustworthy, and runs the Formal Conjectures project to give back.
More from Research
- Test-Time Structure-Space Search lifts antibody design success from 16% to 78% — anshulkundaje · 2026-10-09
- aimotive driving dataset: train labels come from a tracker that sees the future — RexDouglass · 2026-10-09
- Yale-led study finds symbolic structure inside LLM representations — tallinzen · 2026-10-09
- Discrete Diffusion Meetup at COLM: Oct 8, Grand Ballroom — yuntiandeng · 2026-10-09
- Astra 机器人控制实测:Johns Hopkins 深度评测其运动与控制理解 — jmin__cho · 2026-10-09
- Tavily's 93% live-retrieval agent study challenged: prompts were domain-anchored — edwin · 2026-10-09