Text-to-Image Generation on GenEval v2 (test)
58.84Prompt ScoreEPIC
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| EPICExec.=17.0, Call=2.75, kTok in=9.06, kTok out=1.24, Img.=0.79, Temperature=0, Generator=FLUX.2-Klein-4B, Editor=FLUX.2-Klein-4B, MLLM=Qwen3-VL-32B2026.05 | 58.84 | 85.52 | 93.83 | 86.61 | 77.82 | 78.2 | 38.63 | |
| RAISEExec.=22.3, Call=10.3, kTok in=44.41, kTok out=12.30, Img.=8.33, Temperature=0, Generator=FLUX.2-Klein-4B, Editor=FLUX.2-Klein-4B, MLLM=Qwen3-VL-32B2026.05 | 46.32 | 83.49 | 95.2 | 92.53 | 64.98 | 83.02 | 46.01 | |
| FLUX.2-Klein-9BExec.=1.00, Temperature=0, Generator=FLUX.2-Klein-9B2026.05 | 34.16 | 79.06 | 95.82 | 88.86 | 57.55 | 75.13 | 37.55 | |
| BoN+NVILAExec.=32.0, Temperature=0, Generator=FLUX.2-Klein-4B, Editor=FLUX.2-Klein-4B, MLLM=Qwen3-VL-32B2026.05 | 32.75 | 77.09 | 93.25 | 84.93 | 55.46 | 75.15 | 33.59 | |
| T2I-CopilotExec.=1.19, Call=4.37, kTok in=12.16, kTok out=2.54, Img.=1.19, Temperature=0, Generator=FLUX.2-Klein-4B, Editor=FLUX.2-Klein-4B, MLLM=Qwen3-VL-32B2026.05 | 30.38 | 77.94 | 93.76 | 90.05 | 54.62 | 78.23 | 49.03 | |
| FLUX.2-Klein-4BExec.=1.00, Temperature=0, Generator=FLUX.2-Klein-4B2026.05 | 25.11 | 73.07 | 90.92 | 81.73 | 51.91 | 68.22 | 28.01 |