Long-context retrieval
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
100Accuracy
56
Apr 2, 2026
100Accuracy
50
Jul 1, 2026
98.14Retrieval Accuracy (8K)
44
May 22, 2026
62.86nDCG@10
36
Jun 23, 2026
98.67Score
34
Jul 7, 2026
100Accuracy
30
Apr 29, 2026
62.3Biology Score
29
May 8, 2026
100Retrieval Accuracy
29
Jun 2, 2026
10.5Latency (s)
27
Mar 23, 2026
94.67Score
23
Jul 7, 2026
100Accuracy (1024)
22
Jun 2, 2026
24.4Single Score
20
Apr 8, 2026
49.3Single Score
20
Apr 8, 2026
62.57Average Exact Match
20
Mar 12, 2026
4.1Speedup
20
Feb 26, 2026
84.3Accuracy
19
Apr 28, 2026
97.33Score
17
Jul 7, 2026
91.7Accuracy
16
May 11, 2026
95Accuracy
16
May 11, 2026
99.67Score
15
Jul 7, 2026
86.65RULER 32K Retrieval Score
15
Jun 12, 2026
80.1Accuracy
14
Apr 2, 2026
100Accuracy
13
Feb 26, 2026
100Retrieval Accuracy (4K Context)
13
Jun 19, 2026
98.7Success Rate
12
Jul 8, 2026