How DeepMind AlphaProof Nexus Cracks 56-Year-Old Math: Agentic LLM Loops and Lean Formal Verification

Google Deepmind's AlphaProof Nexus solves decades-old math problems for a few hundred dollars

DeepMind's AlphaProof Nexus solved 9 of 353 open Erdős problems — two unsolved for 56 years — at a few hundred dollars each, using Gemini 3.1 Pro to generate Lean proof steps verified by a formal compiler at every iteration. Simple agentic loops grounded by symbolic feedback are converging on specialized trained systems, signaling that LLM + compiler architectures are the next credible pattern for AI agents where logical reliability and auditability are non-negotiable.