List your company

Mathematical Reasoning on AMC (Accuracy, Token count)

97.5AccuracyVanilla
68.976.32583.7591.175Apr 8, 2026

Evaluation results

MethodLinks
VanillaBackbone=Qwen3-14B2026.0497.56,562
NoWAITBackbone=Qwen3-14B2026.0497.55,035
DEERBackbone=Qwen3-14B2026.0497.54,514
DTSRBackbone=Qwen3-14B2026.0497.53,975
VanillaBackbone=Qwen3-32B2026.04956,807
NoWAITBackbone=Qwen3-32B2026.04954,894
DTSRBackbone=Qwen3-32B2026.04954,617
VanillaBackbone=Qwen3-8B2026.0492.57,232
DEERBackbone=Qwen3-32B2026.0492.54,966
NoWAITBackbone=Qwen3-8B2026.04905,448
DTSRBackbone=Qwen3-8B2026.04904,507
DEERBackbone=Qwen3-8B2026.0487.55,155
NoThinkingBackbone=Qwen3-14B2026.0477.52,065
NoThinkingBackbone=Qwen3-8B2026.0472.52,264
NoThinkingBackbone=Qwen3-32B2026.04701,738