List your company

Mathematical Reasoning on AIME 2025 (pass@16)

85Pass@16Ministral 3
20.759237.437154.11570.7929Sep 27, 2025Oct 15, 2025Nov 2, 2025Nov 20, 2025Dec 8, 2025Dec 26, 2025Jan 13, 2026

Evaluation results

MethodLinks
Ministral 3Model Size=14B2026.0185
Qwen3-VLModel Size=8B2026.0179.8
Ministral 3Model Size=8B2026.0178.7
Qwen 3Model Size=14B2026.0173.7
Ministral 3Model Size=3B2026.0172.1
Qwen3-VLModel Size=4B2026.0169.7
GEB-1/πBackbone=Qwen2.5-7B2025.0929.48
GEB-arctanhπBackbone=Qwen2.5-7B2025.0929.38
GEB-πBackbone=Qwen2.5-7B2025.0928.23
DPOBackbone=Qwen2.5-7B2025.0923.23