Open-ended Question Answering
Benchmarks
DatasetSOTA methodMetricTrendResultsLast updated
95.7F1 Score
36
Apr 14, 2026
62.3Accuracy
29
Feb 26, 2026
0.106BLEU-3
27
Feb 26, 2026
73.92Accuracy
22
Feb 26, 2026
74.24Neutral Accuracy
15
Feb 26, 2026
35.1Weighted Recall (Com. Sci.)
12
Feb 26, 2026
0.703ROUGE-1
12
Feb 26, 2026
91Accuracy
11
Mar 9, 2026
77Accuracy
11
Mar 9, 2026
88Accuracy
11
Mar 9, 2026
68.61Kendall's Tau (K)
11
Feb 26, 2026
65.97Kendall's Tau (K)
11
Feb 26, 2026
33.4Problem-level Score
9
May 15, 2026
53.9Problem Level Score
9
May 15, 2026
37.7Subpart AND (v3)
9
May 15, 2026
51.1Subpart-AND (v2)
9
May 15, 2026
96.9Completeness
9
Feb 26, 2026
30.4ROUGE
9
Feb 26, 2026
85.5Accuracy (%)
8
Jun 4, 2026
82.7Exact Match Accuracy
8
Jun 4, 2026
0.656Exact Match Accuracy
8
Jun 4, 2026
32ROUGE-L
8
May 22, 2026
8.58GPT-4o Score
8
Apr 15, 2026
13.91Accuracy
7
May 22, 2026
73.77Accuracy
7
May 22, 2026