Image Generation on ImageNet 256 (val)
2.79Inception ScoreOpen-MAGVIT2-L
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Open-MAGVIT2-Lsingle-NFE=false, external-representation-encoder=false2026.07 | 2.79 | 2.72 | 85.9 | 16.5 | 84 | 36.1 | 114 | 37.6 | 52.9 | 38.4 | 90.6 | 96 | 11.3 | 5.55 | 48.2 | |
| DiT-XL/2single-NFE=false, external-representation-encoder=false2026.07 | 2.11 | 1.35 | 59.9 | 12.7 | 62.2 | 34.5 | 204 | 31.9 | 53.4 | 24.1 | 57.4 | 67.9 | 8 | 1.96 | 44.4 | |
| pMF-H (base)single-NFE=false, external-representation-encoder=false2026.07 | 1.78 | 0.91 | 54.6 | 17.8 | 87.5 | 22.4 | 115 | 30.5 | 65.7 | 31.5 | 70.6 | 94.4 | 10.9 | 8.91 | 43.7 | |
| SiT-XL/2single-NFE=false, external-representation-encoder=false2026.07 | 1.73 | 1.77 | 75.3 | 14.3 | 79 | 35.1 | 258 | 35.3 | 60.9 | 29.5 | 68 | 76.7 | 7.56 | 2.08 | 53.2 | |
| iRDMsingle-NFE=true, external-representation-encoder=true2026.07 | 1.54 | 0.98 | 3.52 | 0.69 | 4.76 | 1.17 | 1.39 | 2.69 | 1.62 | 3.16 | 5.31 | 3.12 | 1.8 | 5.83 | 2.69 | |
| JiT-Hsingle-NFE=true, external-representation-encoder=false2026.07 | 1.26 | 3.06 | 48.1 | 14.2 | 66 | 51.1 | 74.4 | 31.6 | 64.1 | 30.4 | 60.7 | 73.3 | 12.3 | 2.56 | 38.1 | |
| VAR-d30single-NFE=false, external-representation-encoder=false2026.07 | 1.13 | 1.73 | 63.8 | 18.7 | 67.4 | 30.3 | 108 | 34.7 | 61.1 | 29.7 | 62.4 | 75.9 | 11.4 | 1.34 | 40.6 | |
| iMF-XLsingle-NFE=false, external-representation-encoder=false2026.07 | 0.87 | 2.08 | 91 | 17.7 | 98.1 | 40.2 | 594 | 46.4 | 79 | 35 | 92.9 | 93.1 | 10.8 | 3.2 | 86.1 | |
| Drifting-Lsingle-NFE=false, external-representation-encoder=true2026.07 | 0.8 | 3.61 | 136 | 21.1 | 157 | 53 | 845 | 64.2 | 92 | 52.5 | 133 | 128 | 11.8 | 3.98 | 122 | |
| MAR-Hsingle-NFE=false, external-representation-encoder=false2026.07 | 0.79 | 1.19 | 61.5 | 11 | 56.5 | 28.7 | 219 | 30.1 | 57.4 | 20.7 | 65 | 68.1 | 7.18 | 0.37 | 44.8 | |
| LightningDiT-XLsingle-NFE=true, external-representation-encoder=true2026.07 | 0.67 | 0.92 | 36.8 | 8.12 | 41.5 | 21.3 | 61.7 | 19.9 | 39.9 | 14.1 | 41.6 | 50.3 | 5.93 | 1.63 | 24.6 | |
| MDTv2-XL/2single-NFE=false, external-representation-encoder=false2026.07 | 0.63 | 1.23 | 50 | 11.8 | 60.5 | 34.8 | 254 | 31.9 | 59.9 | 19 | 51.3 | 56.8 | 7.82 | 1.69 | 45.8 | |
| SiT-XL/2+REPAsingle-NFE=true, external-representation-encoder=true2026.07 | 0.56 | 1.37 | 47.2 | 11.1 | 55.4 | 29.3 | 171 | 27.7 | 54.1 | 18.4 | 45.3 | 52.5 | 6.4 | 1.37 | 37.3 | |
| DDT-XL/2single-NFE=false, external-representation-encoder=true2026.07 | 0.46 | 1.2 | 49.6 | 11.1 | 57.1 | 33 | 213 | 28.5 | 57.2 | 18.3 | 48.8 | 55.2 | 6.61 | 1.55 | 41.5 | |
| REG-XLsingle-NFE=true, external-representation-encoder=true2026.07 | 0.44 | 1.06 | 33.9 | 8.02 | 46.4 | 24 | 127 | 21.3 | 49.5 | 13.3 | 31.8 | 39.2 | 4.74 | 1.44 | 28.7 | |
| RAE-XLsingle-NFE=true, external-representation-encoder=true2026.07 | 0.36 | 2.34 | 19 | 7.36 | 14.9 | 16.6 | 18.4 | 10.4 | 28.8 | 11.1 | 18.5 | 26.7 | 3.32 | 4.12 | 13 | |
| REPA-E SiT-XL/1single-NFE=true, external-representation-encoder=true2026.07 | 0.34 | 1.26 | 26.1 | 6.19 | 24 | 11.9 | 26.9 | 12.9 | 21.9 | 7.97 | 21.6 | 31.8 | 2.64 | 0.87 | 14 | |
| pMF-H (FD-SIM)single-NFE=true, external-representation-encoder=true2026.07 | 0.22 | 0.36 | 10.3 | 0.37 | 6.34 | 10.5 | 12.5 | 9.39 | 33.2 | 8.72 | 18.5 | 24.5 | 2.28 | 6.56 | 10.3 |