List your company
ResearchFormerly WizWand

Mathematical Reasoning on OlyBench

44.7AccuracyES + TT-MV
2.5813.51524.4535.385Feb 18, 2025Apr 23, 2025Jun 27, 2025Aug 30, 2025Nov 3, 2025Jan 6, 2026Mar 12, 2026
Updated 7mo ago

Evaluation results

MethodLinks
ES + TT-MVModel=OLMo3-7B-Inst2026.0344.7
GPT-4o + QuaSARBackbone=GPT-4o, Prompting Strategy=QuaSAR2025.0244.6
ES + TT-MVModel=OLMo3-7B2026.0343.3
GPT-4o + CoTBackbone=GPT-4o, Prompting Strategy=CoT2025.0241.8
GPT-4o + CoMATBackbone=GPT-4o, Prompting Strategy=CoMAT2025.0240.4
ES + TT-MVModel=Qwen2.5-3B-Inst2026.0339.7
RandOptModel=Qwen2.5-3B-Inst2026.0339.2
Llama-3-70B + QuaSARBackbone=Llama-3-70B, Prompting Strategy=QuaSAR2025.0238.2
ESModel=Qwen2.5-3B-Inst2026.0336.4
Qwen2-72B + QuaSARBackbone=Qwen2-72B, Prompting Strategy=QuaSAR2025.0236.2
RandOptModel=OLMo3-7B-Inst2026.0335.4
PPOModel=Qwen2.5-3B-Inst2026.0334.4
ESModel=OLMo3-7B-Inst2026.0333.1
Qwen2-72B + CoMATBackbone=Qwen2-72B, Prompting Strategy=CoMAT2025.0232.2
RandOptModel=Llama3.1-8B-Inst2026.0332.1
ES + TT-MVModel=Llama3.1-8B-Inst2026.0332.1
Best-of-N‡Model=OLMo3-7B-Inst2026.0330.5
RandOptModel=Qwen2.5-1.5B-Inst2026.0330.4
Qwen2-72B + CoTBackbone=Qwen2-72B, Prompting Strategy=CoT2025.0230.3
ESModel=OLMo3-7B2026.0330.2
GRPOModel=Qwen2.5-3B-Inst2026.0329
ES + TT-MVModel=Qwen2.5-1.5B-Inst2026.0328.9
RandOptModel=OLMo3-7B2026.0328.9
BaseModel=OLMo3-7B-Inst2026.0328.7
Best-of-N‡Model=Qwen2.5-3B-Inst2026.0328
PPOModel=OLMo3-7B-Inst2026.0328
GRPOModel=OLMo3-7B-Inst2026.0327.9
ESModel=Qwen2.5-1.5B-Inst2026.0327.2
TT-MV†Model=OLMo3-7B-Inst2026.0326.5
PPOModel=Qwen2.5-1.5B-Inst2026.0326.3
BaseModel=Qwen2.5-3B-Inst2026.0324.5
TT-MV†Model=Llama3.1-8B-Inst2026.0324.5
GRPOModel=Llama3.1-8B-Inst2026.0323.7
Best-of-N‡Model=Llama3.1-8B-Inst2026.0323
PPOModel=OLMo3-7B2026.0322.9
Llama-3-70B + CoTBackbone=Llama-3-70B, Prompting Strategy=CoT2025.0222.8
ESModel=Llama3.1-8B-Inst2026.0322.3
TT-MV†Model=Qwen2.5-3B-Inst2026.0321.8
Best-of-N‡Model=OLMo3-7B2026.0320.5
Best-of-N‡Model=Qwen2.5-1.5B-Inst2026.0320
BaseModel=Llama3.1-8B-Inst2026.0319.2
GRPOModel=Qwen2.5-1.5B-Inst2026.0318.8
ESModel=Qwen2.5-0.5B-Inst2026.0316.4
PPOModel=Qwen2.5-0.5B-Inst2026.0316.1
PPOModel=Llama3.1-8B-Inst2026.0316.1
TT-MV†Model=Qwen2.5-1.5B-Inst2026.0315.9
RandOptModel=Qwen2.5-0.5B-Inst2026.0315.8
TT-MV†Model=OLMo3-7B2026.0315.7
Qwen2-72BBackbone=Qwen2-72B, Prompting Strategy=Zero-shot2025.0215.6
Llama-3-70BBackbone=Llama-3-70B, Prompting Strategy=Zero-shot2025.0214.6
BaseModel=OLMo3-7B2026.0313.9
ES + TT-MVModel=Qwen2.5-0.5B-Inst2026.0313.5
BaseModel=Qwen2.5-1.5B-Inst2026.0313.4
Best-of-N‡Model=Qwen2.5-0.5B-Inst2026.0310.5
GPT-4oBackbone=GPT-4o, Prompting Strategy=Zero-shot2025.029.9
GRPOModel=Qwen2.5-0.5B-Inst2026.036.9
GRPOModel=OLMo3-7B2026.036.9
TT-MV†Model=Qwen2.5-0.5B-Inst2026.036.7
BaseModel=Qwen2.5-0.5B-Inst2026.034.2