Synthetic Recall on Long-context benchmarks
100Synthetic Recall (8k context)
CLP
Updated 1mo ago
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
2026.03 | 100 | 99.9 | 99.7 | 98.8 | 87.9 | |
2026.03 | 99.8 | 99.8 | 99.4 | 98.2 | 87.1 | |
2026.03 | 99.4 | 99.5 | 99 | 98.4 | 88.6 | |
2026.03 | 99.2 | 99.4 | 95.4 | 52.7 | 40 | |
2026.03 | 99.1 | 99.4 | 97.4 | 40 | 0 | |
2026.03 | 99.1 | 98.5 | 98.7 | 95.4 | 76.4 | |
2026.03 | 99.1 | 98.8 | 97.6 | 74.6 | 29 | |
2026.03 | 98.7 | 97.3 | 96.5 | 93.9 | 79.2 | |
2026.03 | 98.6 | 96.7 | 94 | 71.9 | 43.9 | |
2026.03 | 98.6 | 96.4 | 78.2 | 55.3 | 25.9 | |
2026.03 | 98.6 | 97.1 | 95.9 | 45.1 | 13.9 | |
2026.03 | 98.1 | 96.5 | 91.8 | 88.9 | 69.8 | |
2026.03 | 97.9 | 75 | 29.1 | 13.8 | 6.2 | |
2026.03 | 95.9 | 83.1 | 66.8 | 64.7 | 47.1 | |
2026.03 | 93.6 | 84.4 | 74 | 69.6 | 45.8 | |
2026.03 | 93.5 | 81.1 | 64.8 | 41.2 | 16.3 | |
2026.03 | 90.4 | 77.4 | 60.2 | 54.9 | 31.9 | |
2026.03 | 85.9 | 81.1 | 60.8 | 32.7 | 18.3 | |
2026.03 | 85.4 | 80.1 | 64.5 | 19.4 | 3.4 | |
2026.03 | 85.1 | 80.7 | 52.1 | 33.9 | 21.7 | |
2026.03 | 76.2 | 29.7 | 14.2 | 9.3 | 4.1 |