Science Question Answering
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
96.2Accuracy
7
Mar 16, 2026
75.3Initial Accuracy
6
Jun 23, 2026
37.77Accuracy
6
Jun 9, 2026
51Accuracy
6
Jun 1, 2026
5.2Acceptance Length
6
May 12, 2026
86.5Accuracy
6
Apr 21, 2026
62.3Accuracy
6
Apr 21, 2026
78.59Score
6
Apr 1, 2026
89.04Accuracy
6
Feb 26, 2026
50.9Accuracy (0-shot)
6
Feb 26, 2026
96.3Accuracy
6
Feb 26, 2026
15.65LLMcritic Calls
5
May 11, 2026
69.2Pass@1
5
May 7, 2026
32.32Accuracy
5
Apr 30, 2026
32.83Accuracy
5
Apr 30, 2026
54.7Accuracy (MMLU-Pro OOD Science)
5
Apr 16, 2026
45.9Accuracy
5
Apr 10, 2026
30.2Accuracy
5
Apr 10, 2026
28.01TPS
4
Jun 4, 2026
22.87Normalized Accuracy
4
Jun 2, 2026
2.73SGLang Acceptance Length
4
May 12, 2026
4.04Acceptance Length
4
May 12, 2026
151.7TPS (tok/s)
4
May 7, 2026
78.8Robustness Score (Noise)
4
Apr 24, 2026
78.4Accuracy
4
Apr 14, 2026