GSM8K
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GSM8K (TR)
20.77Accuracy
6
GSM8K ES
10.63Accuracy
6
GSM8K ZH
9.78Accuracy
6
GSM8K IT
25.04Accuracy
6
GSM8K EN
19.71Accuracy
6
GSM8K
83Accg
6
GSM8K
92.89Exact Match Accuracy
6
GSM8K DE
63Compositional Accuracy
6
GSM8K
81.65Accuracy
6
GSM8K sampled 500 items (test)
95Accuracy
6
GSM8K (Tier 1)
89.2SR
6
GSM8K No CoT augmented (test)
0.248Accuracy
6
GSM8K (test)
91.8Accuracy
6
GSM8K (test)
93.3Pass@1
6
GSM8K (test)
70.7Accuracy
6
GSM8K
0.387Inference Cost (USD)
6
GSM8K full (test)
79.8Accuracy
6
GSM8K
66.4Solution Rate
6
GSM8K easy OOD (test)
89.8Accuracy
6
GSM8K (test)
90.2Accuracy
6
GSM8K
80.8pass@1
6
GSM8K
73.8NL-to-Format
6
GSM8K (test)
94.7Accuracy
6
GSM8K
75Exact Match Accuracy (Ngen=128)
6
GSM8K
42.8Average Diffusion Steps
6