Question Answering
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
28.4F1 Score
3
May 8, 2026
6.21JudgeLM Score
3
May 8, 2026
9.7Exclusive Catch Rate (@10%)
3
Apr 29, 2026
31.71Normalized Accuracy
3
Apr 23, 2026
32.48Normalized Accuracy
3
Apr 23, 2026
48.06Token F1
3
Apr 21, 2026
35.3ARC-e Accuracy
3
Apr 15, 2026
31.3ARC-e Accuracy
3
Apr 14, 2026
88.8Accuracy
3
Apr 13, 2026
89.69Accuracy
3
Apr 13, 2026
67Recall
3
Apr 13, 2026
1Missed Abstract Count
3
Apr 13, 2026
3Max References Per Answer Count
3
Apr 13, 2026
40Exact Match
3
Apr 7, 2026
34Exact Match (EM)
3
Apr 7, 2026
83.5Completeness
3
Apr 3, 2026
1,156Average Token Consumption
3
Apr 3, 2026
639Avg Token Consumption
3
Apr 3, 2026
800Average Token Consumption
3
Apr 3, 2026
0.322FC (Mean)
3
Apr 1, 2026
81.12MMLU Accuracy
3
Mar 31, 2026
74.17ACC (Sound)
3
Mar 27, 2026
71.35Accuracy
3
Mar 27, 2026
65.89Direct Perception Accuracy
3
Mar 27, 2026
185.5Score
3
Mar 25, 2026