Development

An internal model takes 91.9% on advanced proof benchmarks

development2026-02-12

Claim

DeepMind unveiled an internal model scoring 91.9% on IMO-ProofBench Advanced, handling doctoral problems in economics and cosmic string physics while autonomously solving four open Erdős problems.

Reported metric: 91.9% / 4 problems

Reporting date: 2026-02-12 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to February 12, 2026 (theinnermostloop.substack.com)

Knowledge graphopen in browser →

Reported in

Welcome to February 12, 2026

Actor

Google DeepMind

Evidence for

Supersedes

Nineteen agents optimize perovskite synthesis in three and a half hours

Referenced by

Cite or correct this record

Citation formatCitation text

An internal model takes 91.9% on advanced proof benchmarks
Innermost Loop KB — secondary record
Reporting date: 2026-02-12
KB permalink: https://nicholsn.github.io/innermost-loop-kb/developments/2026-02-12-deepmind-919-imo-proofbench
Claim: DeepMind unveiled an internal model scoring 91.9% on IMO-ProofBench Advanced, handling doctoral problems in economics and cosmic string physics while autonomously solving four open Erdős problems.
Attributed sources:
- Welcome to February 12, 2026 — https://theinnermostloop.substack.com/p/welcome-to-february-12-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.