List your company
ResearchFormerly WizWand

Mathematical Reasoning on MATH 500 (Acc, Tok #N, CR, Think #R)

96.5AccTA
1.683226.299150.91575.5309Jan 29, 2026Jan 30, 2026Jan 31, 2026Feb 1, 2026Feb 2, 2026Feb 3, 2026Feb 5, 2026
Updated 8mo ago

Evaluation results

MethodLinks
TAModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0196.54,67170100
MI-07Model Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0196.44,61769100
RoutingModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0196.45,4358253.8
DAMI-PrefModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0196.44,3156564
DAMI-ConfModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0196.22,9544434
ThinkingModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0196.16,668100100
MI-05Model Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0196.13,2134872
DAMI-PredModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0195.61,886280
DEERModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0195.46,12592100
MI-03Model Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0195.31,933290
DAMI-PromptModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.01952,8324227
TIESModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0194.81,899280
InstructModel Base Pair=Qwen3-4B-2507-Thinking/Qwen3-4B-2507-Instruct, Sampling Rounds=2x2026.0193.11,74126.10
DAMI-PrefModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0189.43,9526775
DAMI-ConfModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.01893,5266062
DEERModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0188.83,14053100
TAModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0188.82,6704597
MI-07Model Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0188.23,2435595
ThinkingModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.01865,881100100
MI-05Model Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0183.62,3384013
TIESModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0180.2827140
DAMI-PredModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0179.61,823310
MI-03Model Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0176.81,501260
DAMI-PromptModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0176.21,657280
RoutingModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.0163.61,303224.4
InstructModel Base Pair=DeepSeek-R1-Distill-Qwen-7B/Qwen2.5-Math-7B, Sampling Rounds=2x2026.01521,268—0
BRIDGEBackbone=Qwen 2.5-3B2026.0238.2322——
Base ModelBackbone=Qwen 2.5-3B2026.0236.4371——
Instruct ModelBackbone=Qwen 2.5-3B2026.0234.2440——
SuperRLBackbone=Qwen 2.5-3B2026.0233.2355——
Std-CoT KDBackbone=Qwen 2.5-3B2026.0226.6367——
Instruct ModelBackbone=Llama 3.2-3B2026.0219.4181——
BRIDGEBackbone=Llama 3.2-3B2026.0216107——
Mix Distill.Backbone=Qwen 2.5-3B2026.0215.4280——
SuperRLBackbone=Llama 3.2-3B2026.028.6245——
Short-CoTBackbone=Qwen 2.5-3B2026.028.5220——
Std-CoT KDBackbone=Llama 3.2-3B2026.027.2185——
Mix Distill.Backbone=Llama 3.2-3B2026.026.4149——
Short-CoTBackbone=Llama 3.2-3B2026.026.2185——
Base ModelBackbone=Llama 3.2-3B2026.025.33155——