GSM8K
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
GSM8K In-Distribution (test)
91.8Accuracy
5
GSM8k (in-distribution)
91.6Accuracy
5
GSM8K MIP
0.174Accuracy
5
GSM8K
843.05Throughput (TPS)
5
GSM8K sampled 50 (test)
71.5Accuracy
5
GSM8K (test)
44.8Accuracy (5-shot)
5
GSM8K
95.5Accuracy
5
GSM8k (test)
2.14Speedup
5
GSM8K
75.36Score
5
S-PI (GSM8K)
3ASR
4
GSM8K (test)
15.9Query Prefill Latency (ms)
4
GSM8K
81Accuracy (GSM8K)
4
GSM8K
834.52Tokens Per Second (TPS)
4
GSM8K
94.54Accuracy
4
GSM8K (test)
83p@6=0
4
GSM8K (test)
5.16Latency (s)
4
GSM8K
70.58Pass@1
4
GSM8K
72.5Clean Accuracy
4
GSM8K
295.2Inference Time
4
GSM8K
10.67Average Latency
4
GSM8K
72.63S-Format Compliance
4
CAP-GSM8K
21.3RLVR R_temp CE
4
GSM8K
29.75Throughput (TPS)
4
GSM8K
32.44Throughput (TPS)
4
GSM8K (train)
0.381Tokens per Character Ratio (GSM8K train)
4