图像模型排行榜
按独立评测任务比较模型，查看原始成绩与评测配置。
按本项评测的原始 Elo 排名，分数越高越好；不同任务的 Elo 不可直接比较。
| 1 | GPT-Image-2.5 Sunburst OpenAI | 1200.2 Elo 95% 置信区间1189.2–1211.2 | GPT Image 2.5 Sunburst (max) | 18,532 | 2026-11-05 |
| 2 | GPT-Image-2.5 Flare OpenAI | 1183.5 Elo 95% 置信区间1172.5–1194.5 | GPT Image 2.5 Flare (max) | 19,412 | 2026-11-05 |
| 3 | Nano Banana 2.1 Google | 1152.3 Elo 95% 置信区间1139.3–1165.3 | Nano Banana 2.1 | 7,160 | 2026-11-05 |
| 4 | GPT Image 2 OpenAI | 1136.9 Elo 95% 置信区间1124.9–1148.9 | GPT Image 2 (high) | 14,834 | 2026-11-05 |
| 5 | Muse Image Meta | 1128.6 Elo 95% 置信区间1116.6–1140.6 | Muse Image | 13,635 | 2026-11-05 |
| 6 | Qwen-Image-3.0 Pro Alibaba | 1092.6 Elo 95% 置信区间1080.6–1104.6 | Qwen-Image-3.0-Pro | 13,823 | 2026-11-05 |
| 7 | Qwen-Image-3.0 Alibaba | 1070.0 Elo 95% 置信区间1058.0–1082.0 | Qwen-Image-3.0 | 8,341 | 2026-11-05 |
| 8 | Nano Banana 2 Lite Google | 1054.8 Elo 95% 置信区间1042.8–1066.8 | Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) | 8,373 | 2026-11-05 |