Passage Re-ranking on Long-context benchmarks
50.5Performance (8k Context)
FlexPrefill
Updated 1mo ago
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
2026.03 | 50.5 | 30.8 | 19.4 | 9.2 | 5 | |
2026.03 | 50 | 36.3 | 30.1 | 20.4 | 11.2 | |
2026.03 | 49 | 36.7 | 25.9 | 15.6 | 7.5 | |
2026.03 | 48.1 | 32.5 | 25.1 | 18.1 | 8.4 | |
2026.03 | 47.5 | 37.4 | 27.6 | 13.9 | 9.3 | |
2026.03 | 47.3 | 33.2 | 18.1 | 6.8 | 0.6 | |
2026.03 | 46.7 | 27.9 | 17 | 8.7 | 4.3 | |
2026.03 | 46.7 | 32.7 | 24.2 | 14 | 3.8 | |
2026.03 | 46.3 | 37.8 | 32.8 | 20.7 | 2.1 | |
2026.03 | 46.3 | 33.5 | 23.2 | 9.6 | 1.8 | |
2026.03 | 46.3 | 33.2 | 22.7 | 1.5 | 0 | |
2026.03 | 45 | 26 | 14.3 | 9 | 3.4 | |
2026.03 | 44 | 26.8 | 20.1 | 12.5 | 9.6 | |
2026.03 | 42.3 | 32.4 | 24.3 | 12.1 | 7 | |
2026.03 | 42 | 9.4 | 4.5 | 0.3 | 0.2 | |
2026.03 | 41.4 | 31.1 | 12.3 | 8 | 2.2 | |
2026.03 | 34.4 | 20.3 | 8.9 | 0.4 | 0 | |
2026.03 | 33.6 | 20 | 9.2 | 1.7 | 0 | |
2026.03 | 33.5 | 21.6 | 10 | 4.3 | 0 | |
2026.03 | 32.8 | 16.3 | 10 | 1.5 | 1.4 | |
2026.03 | 25.7 | 13 | 8.4 | 3.5 | 2 |