Multi-hop Question Answering
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
29EM
21
Jul 3, 2026
60.61F1
21
Feb 26, 2026
68.8Accuracy
20
Jun 30, 2026
49.6EM
20
Jun 11, 2026
21.7EM
20
Jun 11, 2026
46EM
20
Jun 11, 2026
45.7EM
20
Jun 11, 2026
70.32F1 (56K Context)
20
Jun 2, 2026
59.25Average Score
20
Jun 2, 2026
57.41LLM Accuracy
20
Jun 2, 2026
68.4LLM Accuracy
20
Jun 2, 2026
53.4Exact Match (EM)
20
Apr 21, 2026
24.4Exact Match (EM)
20
Apr 21, 2026
43.6Exact Match (EM)
20
Apr 21, 2026
44.7Exact Match
20
Apr 21, 2026
65.2Token-Level F1
20
Apr 21, 2026
58.7Token-Level F1
20
Apr 21, 2026
62.6Token-level F1
20
Apr 21, 2026
199Inference Time (ms)
20
May 26, 2026
40Exact Match (EM)
20
Mar 17, 2026
42.8EM
20
Mar 10, 2026
60.5EM
20
Mar 10, 2026
56.6EM
20
Feb 26, 2026
0.378EM
20
Feb 26, 2026
22.6Exact Match (EM)
19
Mar 17, 2026