Visual Instruction on VP-Bench
89C2I ScoreFlowInOne
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| FlowInOneEvaluator=Gemini32026.04 | 89 | 70 | 35.5 | 72.7 | 30.2 | 52 | 29.2 | 53.5 | 54 | |
| FlowInOneEvaluator=Qwen3.52026.04 | 85.9 | 72 | 35.4 | 71.3 | 27.2 | 32 | 30.6 | 48.1 | 50.3 | |
| FlowInOneEvaluator=GPT5.22026.04 | 85 | 80 | 7.9 | 50 | 11.6 | 24 | 8.3 | 46.5 | 39.2 | |
| FlowInOneEvaluator=Human2026.04 | 80 | 64.5 | 24.2 | 70.5 | 25.5 | 28 | 25.5 | 40 | 44.9 | |
| Nano Banana GoogleEvaluator=GPT5.22026.04 | 68 | 95.9 | 15.2 | 12.7 | 2.3 | 4 | 13.6 | 30.2 | 30.2 | |
| Nano Banana GoogleEvaluator=Gemini32026.04 | 65 | 98 | 42.3 | 52 | 61.4 | 2 | 54.8 | 72.1 | 56 | |
| Nano Banana GoogleEvaluator=Human2026.04 | 60.2 | 90.4 | 27.1 | 25 | 20 | 5 | 22.9 | 74.2 | 40.6 | |
| Nano Banana GoogleEvaluator=Qwen3.52026.04 | 60 | 95.9 | 38.6 | 36.7 | 25.7 | 4 | 32.1 | 74.4 | 46.9 | |
| Qwen-IE-2509 Wu et al.Evaluator=Qwen3.52026.04 | 27 | 6 | 8 | 8.7 | 4.7 | 4 | 3.3 | 4.7 | 8.3 | |
| Qwen-IE-2509 Wu et al.Evaluator=GPT5.22026.04 | 24 | 12 | 8 | 2 | 2.2 | 6 | 2 | 4.7 | 7.6 | |
| Qwen-IE-2509 Wu et al.Evaluator=Gemini32026.04 | 23 | 4 | 6.9 | 0 | 0 | 2 | 2.3 | 0 | 4.8 | |
| Omnigen2 Wu et al.Evaluator=GPT5.22026.04 | 11 | 2 | 0 | 0 | 0 | 0 | 0 | 2.3 | 1.9 | |
| Kontext Labs et al.Evaluator=GPT5.22026.04 | 9 | 2 | 2.8 | 2 | 0 | 8 | 0.3 | 9.3 | 4.2 | |
| Qwen-IE-2509 Wu et al.Evaluator=Human2026.04 | 6.7 | 0 | 2.9 | 0 | 0 | 0 | 0 | 0 | 1.2 | |
| Kontext Labs et al.Evaluator=Gemini32026.04 | 5 | 2 | 4.8 | 0.7 | 0 | 2 | 1 | 0 | 1.9 | |
| Kontext Labs et al.Evaluator=Qwen3.52026.04 | 5 | 2 | 4.2 | 13.3 | 0 | 6 | 4.7 | 9.3 | 5.6 | |
| Omnigen2 Wu et al.Evaluator=Qwen3.52026.04 | 3 | 2 | 1.7 | 3.4 | 0 | 0 | 0.3 | 4.7 | 1.9 | |
| Omnigen2 Wu et al.Evaluator=Gemini32026.04 | 2 | 2 | 1.7 | 0 | 0 | 0 | 0 | 0 | 0.7 | |
| Omnigen2 Wu et al.Evaluator=Human2026.04 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | |
| Kontext Labs et al.Evaluator=Human2026.04 | 0 | 0 | 4.3 | 0 | 0 | 0 | 0 | 10 | 1.8 |