Benchmark
RISEBench - Visual Plausibility
Scores are percentages.
- Models measured
- 12
- Scores recorded
- 14
- Sources
- 1
- IQ weight
- 1.24
Score distribution
- 87.8%
- Mean
- 91.3%
- Median
- 7.5%
- Std dev
- 73.5%
- 10th percentile
- 94.9%
- 90th percentile
- 72%
- Min
This benchmark has a weight of 1.24 for image-generation models, measured over 10 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | GPT Image 1· RISEBench data | 94.9% |
| 2 | GPT Image 1.5· RISEBench data | 94.9% |
| 3 | GPT Image 2· RISEBench data | 94.9% |
| 4 | Gemini 3 Pro Image· RISEBench data | 94.4% |
| 5 | GPT-Image-1 mini· RISEBench data | 93.7% |
| 6 | Qwen Image Edit2511· RISEBench data | 91.5% |
| 7 | Gemini 2.5 Flash Image· RISEBench data | 91.3% |
| 8 | Doubao Seedream 4-0· RISEBench data | 91.2% |
| 9 | Qwen Image Edit2509· RISEBench data | 86.9% |
| 10 | Flux Kontext Dev· RISEBench data | 85.2% |
| 11 | Gemini 2.0 Flash Experimental· RISEBench data | 82.7% |
| 12 | BAGELw/ CoT· RISEBench data | 81.7% |
| 13 | Step1X-Edit· RISEBench data | 73.5% |
| 14 | BAGEL· RISEBench data | 72% |