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

Question Answering on ARC-Challenge (Accuracy, TFLOPs)

94.37Accuracy

dense alignment

85.488487.794290.192.4058Jun 11, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
94.3721.79
2026.06
94.221.54
2026.06
93.7715.64
2026.06
93.6912.56
2026.06
937.95
2026.06
92.631.16
2026.06
92.631.16
2026.06
92.436.64
2026.06
92.437.42
2026.06
92.3555.5
2026.06
9256.24
2026.06
91.8910.18
2026.06
91.7437.73
2026.06
91.7467.08
2026.06
9119.85
2026.06
9119.85
2026.06
90.9633.91
2026.06
89.8343.48
2026.06
89.397.83
2026.06
89.29.18
2026.06
89.29.18
2026.06
86.783.43
2026.06
86.094.5
2026.06
85.835.01