List your company
ResearchFormerly WizWand

Mathematical Reasoning on MATH500

96.23AccuracyBLASST
24.688443.261761.83580.4083Nov 28, 2025Dec 30, 2025Jan 31, 2026Mar 4, 2026Apr 5, 2026May 7, 2026Jun 9, 2026
Updated 4mo ago

Evaluation results

MethodLinks
BLASSTModel=Qwen3-8B, Sparsity=~50%, Deployment Phase=Decode Phase2025.1296.23
BLASSTModel=Qwen3-8B, Sparsity=~75%, Deployment Phase=Decode Phase2025.1296.07
Dense AttentionModel=Qwen3-8B, Sparsity=Dense, Deployment Phase=Decode Phase2025.1295.87
ReST-MCTSBackbone=Gemma-3-27B2025.1193.2
ReST-MCTSBackbone=GPT-OSS-120B2025.1191.6
APRMBackbone=GPT-OSS-120B2025.1191.4
APRMBackbone=Gemma-3-27B2025.1191.4
Hard Neg.Backbone=GPT-OSS-120B2025.1191.2
ReST-MCTSBackbone=GPT-OSS-20B2025.1191
APRMBackbone=GPT-OSS-20B2025.1191
ARMBackbone=GPT-OSS-120B2025.1189.2
ToTBackbone=Gemma-3-27B2025.1188
ToTBackbone=GPT-OSS-120B2025.1187.6
ToTBackbone=GPT-OSS-20B2025.1187.6
FlowTracerBackbone=Qwen3-8B-Base, Context Length=8K2026.0686.7
LLM-JBackbone=Gemma-3-27B2025.1186.4
Hard Neg.Backbone=Gemma-3-27B2025.1186
ARMBackbone=Gemma-3-27B2025.1186
AttentionBackbone=Qwen3-8B-Base, Context Length=8K2026.0685.9
RandomBackbone=Qwen3-8B-Base, Context Length=8K2026.0685.5
LLM-JBackbone=GPT-OSS-120B2025.1185.2
LLM-JBackbone=GPT-OSS-20B2025.1185.2
GRPOBackbone=Qwen3-8B-Base, Context Length=8K2026.0685.1
CoT-SCBackbone=GPT-OSS-120B2025.1184.4
High-entropyBackbone=Qwen3-8B-Base, Context Length=8K2026.0684.2
GradientBackbone=Qwen3-8B-Base, Context Length=8K2026.0684.2
CorrelationBackbone=Qwen3-8B-Base, Context Length=8K2026.0684.2
FlowTracerBackbone=Qwen3-4B-Base, Context Length=8K2026.0683.1
GradientBackbone=Qwen3-4B-Base, Context Length=8K2026.0682.3
AttentionBackbone=Qwen3-4B-Base, Context Length=8K2026.0682.3
RandomBackbone=Qwen3-4B-Base, Context Length=8K2026.0682
ReST-MCTSBackbone=Gemma-3-12B2025.1181.4
High-entropyBackbone=Qwen3-4B-Base, Context Length=8K2026.0681.2
CorrelationBackbone=Qwen3-4B-Base, Context Length=8K2026.0681.2
GRPOBackbone=Qwen3-4B-Base, Context Length=8K2026.0681
APRMBackbone=Gemma-3-12B2025.1180
FlowTracerBackbone=Qwen3-8B-Base, Context Length=1K2026.0679.7
CoT-SCBackbone=Gemma-3-27B2025.1178.8
AttentionBackbone=Qwen3-8B-Base, Context Length=1K2026.0678.4
High-entropyBackbone=Qwen3-8B-Base, Context Length=1K2026.0678.1
CoT-SCBackbone=GPT-OSS-20B2025.1178
CorrelationBackbone=Qwen3-8B-Base, Context Length=1K2026.0677.9
ToTBackbone=Gemma-3-12B2025.1177.6
RandomBackbone=Qwen3-8B-Base, Context Length=1K2026.0677.4
GradientBackbone=Qwen3-8B-Base, Context Length=1K2026.0677.3
GRPOBackbone=Qwen3-8B-Base, Context Length=1K2026.0677.1
CoT-SCBackbone=Gemma-3-12B2025.1176.8
FlowTracerBackbone=Qwen3-4B-Base, Context Length=1K2026.0675.9
Qwen2.5-Math-7B w/ CADFTBackbone=Qwen2.5-Math-7B, Fine-tuning Protocol=CADFT, Evaluation Protocol=Average@162026.0475.5
AttentionBackbone=Qwen3-4B-Base, Context Length=1K2026.0674.9
High-entropyBackbone=Qwen3-4B-Base, Context Length=1K2026.0674.8
RandomBackbone=Qwen3-4B-Base, Context Length=1K2026.0674.4
CorrelationBackbone=Qwen3-4B-Base, Context Length=1K2026.0674.4
GRPOBackbone=Qwen3-4B-Base, Context Length=1K2026.0674.2
GradientBackbone=Qwen3-4B-Base, Context Length=1K2026.0674.1
BLASSTModel=Llama-3.1-8B, Sparsity=~75%, Deployment Phase=Decode Phase2025.1273.89
BLASSTModel=Llama-3.1-8B, Sparsity=~50%, Deployment Phase=Decode Phase2025.1273.71
Dense AttentionModel=Llama-3.1-8B, Sparsity=Dense, Deployment Phase=Decode Phase2025.1273.4
LLM-JBackbone=Gemma-3-12B2025.1172.8
Qwen2.5-Math-1.5B w/ CADFTBackbone=Qwen2.5-Math-1.5B, Fine-tuning Protocol=CADFT, Evaluation Protocol=Average@162026.0472.3
Ent-RegBase Model=Qwen2.5-14B-Base2026.0571.4
PAPOBase Model=Qwen2.5-14B-Base2026.0571.38
DAPOBase Model=Qwen2.5-14B-Base2026.0570.03
80/20Base Model=Qwen2.5-14B-Base2026.0569.38
Qwen2.5-Math-7B w/ DFTBackbone=Qwen2.5-Math-7B, Fine-tuning Protocol=DFT, Evaluation Protocol=Average@162026.0468.2
Frozen checkpointModel=Moonlight-16B-A3B-Instruct2026.0567.4
Ent-RegBase Model=Qwen2.5-7B-Base2026.0566.9
phi-balancingModel=Moonlight-16B-A3B-Instruct2026.0566.2
PAPOBase Model=Qwen2.5-7B-Base2026.0565.9
80/20Base Model=Qwen2.5-7B-Base2026.0565.67
DAPOBase Model=Qwen2.5-7B-Base2026.0565.57
Qwen2.5-Math-1.5B w/ DFTBackbone=Qwen2.5-Math-1.5B, Fine-tuning Protocol=DFT, Evaluation Protocol=Average@162026.0464.89
ST-MoEModel=Moonlight-16B-A3B-Instruct2026.0564.8
DCCDModel Size=14B2026.0258.6
BaseBase Model=Qwen2.5-14B-Base2026.0556.23
Qwen2.5-Math-7B w/ SFTBackbone=Qwen2.5-Math-7B, Fine-tuning Protocol=SFT, Evaluation Protocol=Average@162026.0453.96
DCCDModel Size=7B2026.0252.8
BaseBase Model=Qwen2.5-7B-Base2026.0550.9
DeepSeekMath-7B w/ CADFTBackbone=DeepSeekMath-7B, Fine-tuning Protocol=CADFT, Evaluation Protocol=Average@162026.0447.8
CDModel Size=14B2026.0247.6
CPModel Size=14B2026.0247
DCCDModel Size=3B2026.0246.8
CFModel Size=14B2026.0245.8
CFModel Size=7B2026.0244.2
Qwen2.5-Math-1.5B w/ SFTBackbone=Qwen2.5-Math-1.5B, Fine-tuning Protocol=SFT, Evaluation Protocol=Average@162026.0443.76
CDModel Size=7B2026.0243.6
DeepSeekMath-7B w/ DFTBackbone=DeepSeekMath-7B, Fine-tuning Protocol=DFT, Evaluation Protocol=Average@162026.0441.46
CPModel Size=7B2026.0240.4
Qwen2.5-Math-7BBackbone=Qwen2.5-Math-7B, Fine-tuning Protocol=None, Evaluation Protocol=Average@162026.0440.12
DCCDModel Size=1.5B2026.0238.2
DCCDModel Size=8B2026.0235
CDModel Size=3B2026.0233.4
CFModel Size=3B2026.0232.4
Qwen2.5-Math-1.5BBackbone=Qwen2.5-Math-1.5B, Fine-tuning Protocol=None, Evaluation Protocol=Average@162026.0431.66
LLaMA-3.1-8B w/ CADFTBackbone=LLaMA-3.1-8B, Fine-tuning Protocol=CADFT, Evaluation Protocol=Average@162026.0431.2
S³Backbone=LLaDA-8B-Instruct, Decoding Strategy=S³, N (S³)=4, b (S³)=2, K (Best-of-K)=642026.0430.2
CPModel Size=3B2026.0230
CDModel Size=8B2026.0228.6
Best-of-KBackbone=LLaDA-8B-Instruct, Decoding Strategy=Best-of-K, K (Best-of-K)=82026.0428.2
LLaMA-3.1-8B w/ DFTBackbone=LLaMA-3.1-8B, Fine-tuning Protocol=DFT, Evaluation Protocol=Average@162026.0427.44