DeepMind's AlphaProof Nexus autonomously solves 9 open Erdős problems, two unsolved for 56 years

Dr_Singularity · x · 2026-10-10

Google DeepMind's AlphaProof Nexus has delivered a major math research breakthrough:

The key technique: pairing an LLM's proof generation with Lean's formal verification — the AI proposes a proof, the checker validates every logical step, and failed attempts generate feedback for retry. Only proofs passing the checker are accepted. Researchers also benchmarked a simpler baseline: repeatedly generating candidate proofs with an LLM and checking them in Lean, alongside the more sophisticated search framework.

Related event: DeepMind's AlphaProof Nexus Lands in Science, Solving 9 Open Erdős Problems(7 posts)→

Original post →

More from Research

Research channel →