AIMAGENTA reportedly solved all six IMO 2026 problems with K2-Horizon-7B
An announcement introducing MAGENTA says it combines natural-language reasoning with formal verification and achieved 100% accuracy across 93 math competition problems using K2-Horizon models of different sizes.
TLDR
The MAGENTA announcement describes a feedback loop between natural-language reasoning and formal verification. It reports 100% accuracy across 93 problems from AIME 2025, AIME 2026 and HMMT February 2026 using K2-Horizon models of different sizes. With the compact K2-Horizon-7B, MAGENTA reportedly solved all six IMO 2026 problems. The announcement says those solutions were assessed using automated grading with reference answers and separately reviewed by IMO medalists.
Combined views
6.9K
1 Source, first seen 1d ago
MAGENTA reportedly solved all six IMO 2026 problems with K2-Horizon-7B
An announcement introducing MAGENTA says it combines natural-language reasoning with formal verification and achieved 100% accuracy across 93 math competition problems using K2-Horizon models of different sizes.