From Failing 2+2 to PhD-Level Math: A Timeline of AI's Three-Year Leap

haider1 · x · 2026-09-09

A viral post sums up AI's math progress year by year: in 2023 models confidently claimed 2+2=5; by 2024 they handled school math but struggled with harder high-school problems; in 2025 they won IMO gold and started tackling undergrad-level math; and in 2026 they contribute like math PhDs — first through literature-heavy results, now finding genuinely novel solutions to open problems.

Original post →

More from Models

Models channel →