List your company
ResearchFormerly WizWand

Mathematical Reasoning on HMMT (Pass@1)

73.3Pass@1Pairwise
17.6632.10546.5560.995Jan 20, 2026Feb 16, 2026Mar 16, 2026Apr 13, 2026May 11, 2026Jun 8, 2026Jul 6, 2026
Updated 3mo ago

Evaluation results

MethodLinks
PairwiseVerifier=V1, Verification Budget=3N budget, Sampling Strategy=Best-of-N2026.0773.3
LLM-as-a-VerifierVerification Budget=2N budget, Sampling Strategy=Best-of-N2026.0773.3
EvoEnvBase Model=Nemotron-Cascade-8B2026.0564.2
PointwiseMechanism=LM judge, Verification Budget=N budget, Sampling Strategy=Best-of-N2026.0763.5
UntrainedBase Model=Nemotron-Cascade-8B2026.0561.5
EvoEnvBase Model=Qwen3-4B-Thinking-25072026.0560
UntrainedBase Model=Qwen3-4B-Thinking-25072026.0556.9
R-ZeroBase Model=Nemotron-Cascade-8B2026.0555.8
RLVEBase Model=Nemotron-Cascade-8B2026.0555.4
R-ZeroBase Model=Qwen3-4B-Thinking-25072026.0553.5
Base modelSampling Strategy=Best-of-N2026.0752
RLVEBase Model=Qwen3-4B-Thinking-25072026.0551.5
DAPOBase Model=Qwen3-4B-Thinking-25072026.0548.8
DAPOBase Model=Nemotron-Cascade-8B2026.0540
UntrainedBase Model=Qwen3-4B-Instruct-25072026.0530
EvoEnvBase Model=Qwen3-4B-Instruct-25072026.0530
RLVEBase Model=Qwen3-4B-Instruct-25072026.0527.7
R-ZeroBase Model=Qwen3-4B-Instruct-25072026.0526.5
e3-1.7B + InT+ RLRL Data Size=64, Evaluation Protocol=8 rollouts avg.2026.0124.58
e3-1.7B + RLRL Data Size=1216, Evaluation Protocol=8 rollouts avg.2026.0122.5
e3-1.7B + Distill + RLRL Data Size=1216, Evaluation Protocol=8 rollouts avg.2026.0122.5
DAPOBase Model=Qwen3-4B-Instruct-25072026.0519.8