Flux.1
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Text-to-Image Generation | FLUX.1 (dev) | Image Reward1.0527 | 56 | |
| Text-to-Image Generation | FLUX.1 1024x1024 resolution (dev) | ImageReward1.095 | 20 | |
| Image Generation | FLUX-1 (dev) | Image Reward1.1998 | 20 | |
| Text-to-image generation | FLUX.1 200 prompts (dev) | FLOPs (T)595.12 | 15 | |
| Text-to-Image Alignment | FLUX.1 dev v1.0 (test) | HPS-v2.10.372 | 13 | |
| Text-to-Image Generation | FLUX.1 1.0 (dev) | FLOPs (T)388.27 | 12 | |
| Latent Decoding | FLUX.1 (dev) | MUSIQ (paq2piq)73.4 | 10 | |
| Text-to-image generation | FLUX.1 DrawBench (dev) | PSNR26.285 | 10 | |
| Text-to-Image Generation | FLUX.1 T=50 (dev) | GenEval64.61 | 7 | |
| Human preference alignment | FLUX.1 (dev) | HPS-v2.10.36 | 6 | |
| Text-to-Image Synthesis | FLUX.1 dev (test) | Reward1.45 | 5 | |
| Text-to-Image Generation | Flux.1 1024 x 1024 (dev) | HPSv2.130.32 | 5 | |
| Image Generation | FLUX.1 schnell | Time (s)0.94 | 5 | |
| Per-step latency | FLUX.1 Schnell | Per-step Latency (ms)432 | 4 | |
| 3D Layout-to-Image Generation | Flux.1 layout dataset v1 (test) | CLIP Score (T2I)0.333 | 4 | |
| Text-to-Image Generation | FLUX.1 dev 12B (test) | Image Reward1.258 | 3 | |
| AI-generated image detection | FLUX.1 | Accuracy96.2 | 3 |
Showing 17 of 17 rows