ARC
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ARC Challenge
84.61Accuracy
16
ARC-E
87.8Accuracy
16
ARC
1.7Average Rank
16
ARC Challenge Italian
93.5Flexible Extract
16
ARC Challenge
95.1Exact Match Accuracy
16
ARC Challenge
27.22Accuracy
16
ARC Challenge v1 (test)
56.7Normalized Accuracy
16
ARC Easy v1 (test)
83.5Accuracy
16
ARC-IT
95Accuracy
16
ARC Challenge
0.636PR-AUC
16
ARC
92.11ARC Easy Accuracy
16
ARC Challenge
95.3Accuracy
16
ARC Challenge o.o.d.
90.65Accuracy
15
ARC Challenge
97.7Accuracy (ARC)
15
ARC-C
81.23Pass@1
15
ARC Easy
92.18Accuracy (ARC-E)
15
ARC-E (test)
79.5Accuracy
15
ARC Easy
0.096Mean Regret
15
ARC Challenge
0.095Mean Per-Step Regret
15
ARC-c
84.6Pass@1
15
ARC Challenge (test)
84.04Accuracy
15
ARC-C (test)
79.9Accuracy
15
ARC-C French (test)
90Accuracy
15
ARC Challenge
67.1EN Accuracy
14
ARC Easy
89.7EN Accuracy
14