List your company

Mathematical Reasoning on MATH-500 1.0 (test)

86.6AccuracyMUR
42.9254.2665.676.94Jul 20, 2025

Evaluation results

MethodLinks
MURBackbone=Qwen3-30B-A3B, Search Strategy=LLM as a critic2025.0786.61,594
MURBackbone=Qwen3-30B-A3B, Search Strategy=φ-Decoding2025.07861,700
SMARTBackbone=Qwen3-30B-A3B, Search Strategy=Guided search2025.0785.45,601
Per-Step ScaleBackbone=Qwen3-30B-A3B, Search Strategy=Guided search2025.0785.27,345
Per-Step ScaleBackbone=Qwen3-30B-A3B, Search Strategy=LLM as a critic2025.0784.82,985
MURBackbone=Qwen3-30B-A3B, Search Strategy=Guided search2025.0784.61,564
Avg uncertaintyBackbone=Qwen3-30B-A3B, Search Strategy=LLM as a critic2025.0784.42,502
Avg uncertaintyBackbone=Qwen3-30B-A3B, Search Strategy=Guided search2025.0784.23,779
Vanilla CoTBackbone=Qwen3-30B-A3B2025.0783.4606
SMARTBackbone=Qwen3-30B-A3B, Search Strategy=LLM as a critic2025.0783.41,057
SMARTBackbone=Qwen3-30B-A3B, Search Strategy=φ-Decoding2025.0782.83,052
Per-Step ScaleBackbone=Qwen3-30B-A3B, Search Strategy=φ-Decoding2025.0782.63,609
Avg uncertaintyBackbone=Qwen3-30B-A3B, Search Strategy=φ-Decoding2025.0781.21,990
Per-Step ScaleBackbone=GLM-4-9B, Search Strategy=Guided search2025.0756.63,494
Avg uncertaintyBackbone=GLM-4-9B, Search Strategy=LLM as a critic2025.07561,008
SMARTBackbone=GLM-4-9B, Search Strategy=Guided search2025.0753.42,304
Per-Step ScaleBackbone=GLM-4-9B, Search Strategy=LLM as a critic2025.0753.21,037
MURBackbone=GLM-4-9B, Search Strategy=LLM as a critic2025.0753.2746
MURBackbone=GLM-4-9B, Search Strategy=Guided search2025.0749.6901
Avg uncertaintyBackbone=GLM-4-9B, Search Strategy=Guided search2025.0748.61,196
SMARTBackbone=GLM-4-9B, Search Strategy=LLM as a critic2025.0748.6579
MURBackbone=GLM-4-9B, Search Strategy=φ-Decoding2025.07481,531
Avg uncertaintyBackbone=GLM-4-9B, Search Strategy=φ-Decoding2025.0747.81,316
SMARTBackbone=GLM-4-9B, Search Strategy=φ-Decoding2025.07472,386
Vanilla CoTBackbone=GLM-4-9B2025.0745.2473
Per-Step ScaleBackbone=GLM-4-9B, Search Strategy=φ-Decoding2025.0744.63,817