General Capability on OLMES benchmarks
51.4Average ScoreFull-IT
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Full-ITK=32, BoS=false2026.05 | 51.4 | 44.9 | 57.8 | 74.6 | 46.6 | 36.1 | |
| IT-SPEED-28+BoSK=28, BoS=true2026.05 | 51.3 | 45.7 | 57 | 75.2 | 46.6 | 34.9 | |
| IT-SPEED-24+BoSK=24, BoS=true2026.05 | 51.2 | 46 | 58 | 75.4 | 45.3 | 33.9 | |
| IT-SPEED-28K=28, BoS=false2026.05 | 50.2 | 44 | 56.1 | 73.8 | 45.2 | 35 | |
| IT-SPEED-20+BoSK=20, BoS=true2026.05 | 49.9 | 45.1 | 56.1 | 75.5 | 42.9 | 32.3 | |
| IT-SPEED-24K=24, BoS=false2026.05 | 49.1 | 41.7 | 54.8 | 75.6 | 43.1 | 34.3 | |
| IT-SPEED-20K=20, BoS=false2026.05 | 48.6 | 42.3 | 55.6 | 73.6 | 42.8 | 32 | |
| IT-SPEED-16+BoSK=16, BoS=true2026.05 | 45.4 | 43 | 44 | 71 | 40.6 | 29.9 | |
| IT-SPEED-16K=16, BoS=false2026.05 | 44.3 | 39.3 | 45.6 | 74.1 | 36.1 | 29.1 |