List your company

Mathematical Reasoning on OlympiadBench (Acc, Tok, CR)

81AccuracyVanilla
53.9660.986875.02Apr 19, 2026

Evaluation results

MethodLinks
VanillaBackbone=Qwen3-8B2026.048110,518100
TRACEBackbone=Qwen3-8B2026.04808,08876.9
VanillaBackbone=Qwen3-4B2026.0477.910,294100
TRACEBackbone=Qwen3-4B2026.0477.69,48992.2
VanillaBackbone=Gemini2.5-Flash2026.0477.66,684100
DynasorBackbone=Qwen3-8B2026.0477.39,14887
TRACEBackbone=Gemini2.5-Flash2026.0475.34,29164.2
DEERBackbone=Qwen3-4B2026.0474.79,69694.2
DEERBackbone=Qwen3-8B2026.0474.28,07476.8
NoThinkBackbone=Gemini2.5-Flash2026.0474.16,29694.2
DynasorBackbone=Qwen3-4B2026.0473.89,13388.7
TALEBackbone=Qwen3-8B2026.0473.18,41780
DynasorBackbone=Gemini2.5-Flash2026.0472.75,91988.5
DEERBackbone=Gemini2.5-Flash2026.0470.35,80586.8
TALEBackbone=Qwen3-4B2026.0469.58,52182.8
TALEBackbone=Gemini2.5-Flash2026.0468.44,95074
VanillaBackbone=R1-Distilled-Llama-8B2026.0458.26,286100
TRACEBackbone=R1-Distilled-Llama-8B2026.0457.54,78476.1
NoThinkBackbone=Qwen3-8B2026.0457.21,71016.3
DynasorBackbone=R1-Distilled-Llama-8B2026.0456.95,81492.5
TALEBackbone=R1-Distilled-Llama-8B2026.0456.25,71390.9
NoThinkBackbone=R1-Distilled-Llama-8B2026.04566,15097.8
DEERBackbone=R1-Distilled-Llama-8B2026.0455.45,44386.6
NoThinkBackbone=Qwen3-4B2026.04551,80417.5