Zero-shot Evaluation
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
76Accuracy
278
Feb 26, 2026
72.79Mean Accuracy (Zero-shot)
223
Jun 25, 2026
77.62MMLU
218
Feb 26, 2026
Arc-Challenge, Arc-Easy, HellaSwag, LAMBADA-openai, LAMBADA-standard, PIQA, and WinoGrande Zero-Shot
76.71Accuracy (0-shot)
123
Jun 12, 2026
63.54Accuracy
112
May 26, 2026
77.07Average Accuracy
102
Jul 9, 2026
68.05Accuracy
95
Jul 9, 2026
73.81Average Accuracy (Zero-shot)
85
May 27, 2026
73.39ARC-C
85
May 20, 2026
59.3ARC-c Accuracy
82
Jun 4, 2026
79.95Avg. Accuracy
74
May 20, 2026
80.05Average Accuracy
70
May 19, 2026
76Average Accuracy
61
Feb 26, 2026
68.6Average Accuracy
60
Jun 11, 2026
71.08Mean Accuracy
59
Mar 10, 2026
72.9Accuracy
57
Apr 7, 2026
71.76Average Accuracy
55
Jun 4, 2026
79.95Average Accuracy
38
Feb 26, 2026
73.330-shot Average Score
33
Jul 9, 2026
84.4PiQA Accuracy (Zero-shot)
30
Apr 10, 2026
60.31Accuracy (Zero-shot)
29
Apr 30, 2026
77.28Average Score
29
Jun 5, 2026
73.99Task Avg Score
26
May 19, 2026
52.1ARC-C Accuracy
26
Mar 10, 2026
76.57mAcc
25
Feb 26, 2026