Share your thoughts, 1 month free Claude Pro on usSee more

Mathematical Reasoning on GSM8k (Accuracy)

100Accuracy

Phi-4 pass@N (Upper Bound)

Updated 14d ago

Evaluation Results

Method	Links
Phi-4 pass@N (Upper Bound) 2025.11		100
Majority Voting 2025.11		100
H4-Qwen2.5-PRM-1.5B-0.2 2025.11		100
Universal-PRM-Qwen2.5-Math-7B 2025.11		100
Qwen2.5-Math-7B-PRM800k 2025.11		100
Qwen2.5-Math-7B-PRM 2025.11		100
ReProbe, Attn+Logit, Qwen3-8B-anno 2025.11		100
Skywork-PRM-1.5B 2025.11		99.5
RLHFlow-PRM-Deepseek-Data 2025.11		98.9
Math-Shepherd-PRM-7B 2025.11		98.4
RLHFlow-PRM-Mistral-Data 2025.11		98.4
Phi-4 pass@1 (Lower Bound) 2025.11		97.9
SQ-format 2025.12		96.21
DeepSeek-R1 (BF16) 2025.12		95.83
ZEDA 2026.05		95.5
Qwen3-30B-A3B 2026.05		95.4
NETSFT→OPD 2026.05		95.4
Dynamic Skipping 2026.05		95.2
GLM-4.7-Flash 2026.05		95.2
ZEDASFT 2026.05		95.2
ZEDASFT 2026.05		94.8
NETSFT 2026.05		94.7
ZEDA 2026.05		94.4
NETSFT→OPD 2026.05		94.2
NETSFT 2026.05		94.1
AdaMoE 2026.05		93.9
Dynamic Skipping 2026.05		93.8
SRQ-driven steering 2026.04		93.3
SRQ-driven steering 2026.04		93.1
PPoT + Qwen2.5 Coder 2026.04		93.02
DS-R1-Distill-Qwen-7B-ScaleQuest 2024.10		93
SRQ-driven steering 2026.04		93
AUTOIF (LLaMA-3-70B w/ Online DPO) 2024.06		92.7
LLaMA3-70B-Instruct 2024.06		92.6
Conceptor 2026.04		92.6
PCA-CAA 2026.04		92.4
AdaMoE 2026.05		92.4
CAA 2026.04		92.2
GPT-4 2024.03		92
Base LLM 2026.04		91.9
Vanilla 2026.01		91.66
DS-R1-Distill-Qwen-7B 2024.10		91.5
Ans. Aug 2026.01		91.5
Qwen2.5 Coder 2026.04		91.2
SAT 2026.01		91
BF16 2025.12		90.45
ES + TT-MV 2026.03		90.2
SRQ-driven steering 2026.04		89.9
SQ-format 2025.12		89.84
SRQ-driven steering 2026.04		89.6
RandOpt 2026.03		89.5
SRQ-driven steering 2026.04		89.5
ES 2026.03		89.1
PCA-CAA 2026.04		89.1
PPoT + Qwen2.5 Coder 2026.04		89.08
Know. Aug 2026.01		88.8
SQ-format 2025.12		88.78
Conceptor 2026.04		88.6
Qwen3-8B 2026.01		88.42
PPO 2026.03		88.4
CAA 2026.04		88.4
SRQ-driven steering 2026.04		88.4
AUTOIF (Qwen2-72B-Instruct w/ Online DPO) 2024.06		88.2
SRQ-driven steering 2026.04		88.2
SQ-format 2025.12		88.17
SRQ-driven steering 2026.04		88.1
Qwen2-72B-Instruct 2024.06		87.9
ES + TT-MV 2026.03		87.9
Base LLM 2026.04		87.8
SDAR-4B-Chat 2026.01		87.64
ES + TT-MV 2026.03		87.5
PCA-CAA 2026.04		87.5
SDAR-4B-Chat 2026.01		87.41
ES 2026.03		87.2
VIST2-8B 2026.01		87.15
RandOpt 2026.03		87.1
Conceptor 2026.04		87.1
GRPO 2026.03		87
RandOpt 2026.03		86.7
CAA 2026.04		86.7
SDAR-4B-Chat 2026.01		86.58
ES + TT-MV 2026.03		86.4
Qwen2.5 Coder 2026.04		86.2
Qwen3-4B 2026.01		85.97
ES 2026.03		85.8
Base LLM 2026.04		85.8
FP16 2025.09		85.1
Best-of-N‡ 2026.03		85
RandOpt 2026.03		85
RTN 2025.09		84.8
PPoT + Qwen2.5 Coder 2026.04		84.76
RTN 2025.09		84.7
GRPO (GT) 2026.01		84.61
GPTQ 2025.09		84.5
GPTQ 2025.09		84.4
PRISM 2026.01		84.31
VIST2-4B 2026.01		84.31
Qwen2.5-14B 2026.04		84
Bit-by-Bit 2026.04		84
Best-of-N‡ 2026.03		83.5

Showing 100 of 388 rows