Our new X account is live! Follow @wizwand_team for updates

Mathematical Reasoning on Minerva Math

63.2Accuracy

Pass@8 (Upper Bound)

Updated 4d ago

Evaluation Results

Method	Links
Pass@8 (Upper Bound) 2026.01		63.2
Pass@8 (Upper Bound) 2026.01		57.7
Qwen2.5-Math-PRM-7B 2026.01		54
Majority Vote@8 2026.01		53.3
Qwen2.5-Math-7B-NAIT 2026.01		52.7
EurusPRM-Stage1 2026.01		52.6
EurusPRM-Stage2 2026.01		51.5
Skywork-PRM-Qwen2.5-7B 2026.01		51.1
Qwen2.5-Math-7B-MCRD 2026.01		50.8
RLHFlow-PRM-Mistral-8B 2026.01		50.4
RLHFlow-PRM-DeepSeek-8B 2026.01		50
Math-Shepherd-PRM-7B 2026.01		49.6
Qwen2.5-Math-7B-MC 2026.01		49.3
Greedy 2026.01		49.2
Skywork-PRM-Qwen2.5-7B 2026.01		48.5
Pass@8 (Upper Bound) 2025.05		46.7
Majority Vote@8 2026.01		46.7
Qwen2.5-Math-PRM-7B 2026.01		46.7
Qwen3-235B-A22B 2025.12		46.69
Pass@8 (Upper Bound) 2026.01		46
Qwen2.5-Math-7B-NAIT 2026.01		45.8
EurusPRM-Stage2 2026.01		45.6
Qwen3-32B 2025.12		45.22
Qwen2.5-Math-7B-MCRD 2026.01		45.1
Math-Shepherd-PRM-7B 2026.01		44.8
RLHFlow-PRM-DeepSeek-8B 2026.01		44.8
EurusPRM-Stage1 2026.01		44.8
FIGR 2025.12		44.49
Qwen2.5-Math-7B-MC 2026.01		44.2
Reasoning Palette (Linear Decay) 2025.12		43.77
Text-only RL 2025.12		43.38
Greedy 2026.01		43
Reasoning Palette (Linear Decay) 2025.12		42.67
Reasoning Palette (Two-Phase) 2025.12		42.57
Reasoning Palette (Linear Decay) 2025.12		42.53
RLHFlow-PRM-Mistral-8B 2026.01		42.3
Reasoning Palette (Two-Phase) 2025.12		42.2
Reasoning Palette (Two-Phase) 2025.12		41.37
Qwen3-VL-32B-Instruct 2025.12		41.18
Reasoning Palette (Linear Decay) 2025.12		40.91
Majority @8 2025.05		40.1
Qwen3-VL-8B-Instruct 2025.12		40.07
RLOO 2025.12		39.48
GRPO 2025.12		39.39
Qwen3-32B 2025.12		39.34
Skywork-PRM-7B 2025.05		39.1
GRPO 2025.12		39.07
Reasoning Palette (Two-Phase) 2025.12		38.75
Majority Vote@8 2026.01		38.6
GLM-4.5V 2025.12		38.6
RLHFlow-PRM-Deepseek-8B 2025.05		38.2
Skywork-PRM-1.5B 2025.05		38.2
SCOPE 2025.05		38.2
SCOPE (w/o AST normalization) 2025.05		38.2
Skywork-PRM-Qwen2.5-7B 2026.01		38.2
Qwen2.5-Math-7B-PRM800K 2025.05		38.1
EurusPRM-Stage1 2025.05		37.9
PRISM 2026.01		37.9
EurusPRM-Stage2 2025.05		37.7
RLHFlow-PRM-Mistral-8B 2025.05		37.5
Qwen2.5-Math-PRM-7B 2026.01		37.5
SCOPE (Step Replacement) 2025.05		37.1
SCOPE (Step Skipping) 2025.05		37.1
Qwen2.5-Math-7B-NAIT 2026.01		37.1
LLaDA RCD 2026.01		37
INT. 2026.01		36.8
Qwen2.5-Math-7B-MCRD 2026.01		36.7
SCOPE (w/o code translation) 2025.05		36.4
RLHFlow-PRM-Mistral-8B 2026.01		36.4
EurusPRM-Stage1 2026.01		36.4
RLOO 2025.12		36.07
EurusPRM-Stage2 2026.01		35.6
RLHFlow-PRM-DeepSeek-8B 2026.01		34.9
Greedy 2025.05		34.6
GRPO (GT) 2026.01		34.2
Math-Shepherd-PRM-7B 2025.05		33.8
Math-Shepherd-PRM-7B 2026.01		33.8
Qwen2.5-Math-7B-MC 2026.01		33.5
LLaDA Base 2026.01		31.4
LLaDA SeqD 2026.01		31.1
PRISM 2026.01		29.4
Greedy 2026.01		27.2
GRPO (GT) 2026.01		25.7
Qwen3-4B + TTSV 2025.12		25.4
Qwen3-4B 2025.12		24.6
PRM 2026.01		20.2
Reasoning Palette (Linear Decay) 2025.12		13.98
RLOO 2025.12		13.05
Reasoning Palette (Linear Decay) 2025.12		12.82
Reasoning Palette (Two-Phase) 2025.12		12.59
GRPO 2025.12		12.36
Base 2026.01		12.1
Reasoning Palette (Two-Phase) 2025.12		11.99
Base 2026.01		9.6
Bagel-7B-MoT 2025.12		9.56
DeepEyes 2025.12		5.15
Bagel-Zebra-CoT 2025.12		2.94
Chain-of-Focus 2025.12		0.74
INT. 2026.01		0.4
PRM 2026.01		0.4