Benchmark
GEditBench v2 - Instruction Following
- Models measured
- 12
- Scores recorded
- 14
- Sources
- 1
- IQ weight
- 1.27
Score distribution
- 1,001.7
- Mean
- 1,033
- Median
- 134.3
- Std dev
- 807
- 10th percentile
- 1,126
- 90th percentile
- 787
- Min
This benchmark has a weight of 1.27 for image-generation models, measured over 12 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 1.5· GEditBench v2 | 1,260 |
| 2 | Gemini 3 Pro Image· GEditBench v2 | 1,126 |
| 3 | Doubao Seedream 4.5· GEditBench v2 | 1,111 |
| 4 | Qwen Image Edit2511· GEditBench v2 | 1,095 |
| 5 | FLUX.2 [klein] 9B· GEditBench v2 | 1,083 |
| 6 | FLUX.2 [dev]· GEditBench v2 | 1,037 |
| 7 | Qwen Image Edit2509· GEditBench v2 | 1,033 |
| 8 | Longcat Image Edit· GEditBench v2 | 1,018 |
| 9 | FLUX.2 [klein] 4B· GEditBench v2 | 1,007 |
| 10 | Qwen Image Edit· GEditBench v2 | 991 |
| 11 | Flux Kontext Dev· GEditBench v2 | 849 |
| 12 | BAGEL· GEditBench v2 | 820 |
| 13 | omnigen-v2· GEditBench v2 | 807 |
| 14 | GLM-Image· GEditBench v2 | 787 |