List your company

Mathematical Problem Solving on AIME 2025 (Accuracy avg@k)

73.54Accuracy (avg@k)TRACE (FKL on Ky)
15.258430.389245.5260.6508May 11, 2026

Evaluation results

MethodLinks
TRACE (FKL on Ky)Backbone=Qwen3-8B, Symmetric Think evaluation=true2026.0573.54
TRACE (RKL on Ey)Backbone=Qwen3-8B, Symmetric Think evaluation=true2026.0570.42
GRPOBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.0568.96
RLSDBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.0568.54
BaseBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.0567.5
SRPOBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.0554.17
SDPOBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.0541.25
TRACE (RKL on Ey)Backbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.0538.33
TRACE (FKL on Ky)Backbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.0536.3
BaseBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.0535
GRPOBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.0535
RLSDBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.0532.92
SRPOBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.0523.75
SDPOBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.0517.5