Long-context memory management
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
68.5Precision
18
Jun 11, 2026
31.3Precision
18
Jun 11, 2026
83Single-Doc Recall
16
Jun 25, 2026
1,294Added Tokens (mean)
4
Mar 23, 2026
52Overall QA F1
2
Apr 1, 2026