List your company

Math Reasoning on AIME 2024, AIME 2025, and HMMT 2025 Aggregate

66.9Macro Avg. AccuracyPAINT
34.55642.95351.3559.747Apr 29, 2026

Evaluation results

MethodLinks
PAINTBackbone=Qwen3-8B Instruct2026.0466.9
OPSDBackbone=Qwen3-8B Instruct2026.0464.8
PAINTBackbone=Qwen3-4B Instruct2026.0464.4
GRPOBackbone=Qwen3-8B Instruct2026.0464
OPSDBackbone=Qwen3-4B Instruct2026.0463.6
GRPOBackbone=Qwen3-4B Instruct2026.0462.7
No updateBackbone=Qwen3-8B Instruct2026.0461.8
No updateBackbone=Qwen3-4B Instruct2026.0461.2
SFTBackbone=Qwen3-8B Instruct2026.0459.8
SFTBackbone=Qwen3-4B Instruct2026.0458.6
PAINTBackbone=Qwen3-1.7B Instruct2026.0444.5
OPSDBackbone=Qwen3-1.7B Instruct2026.0443.4
GRPOBackbone=Qwen3-1.7B Instruct2026.0437.7
No updateBackbone=Qwen3-1.7B Instruct2026.0437.1
SFTBackbone=Qwen3-1.7B Instruct2026.0435.8