EditReward-Compass - Visual Quality: leaderboard
Accuracy (%) of multimodal LLMs used as reward models on the preference pairs that differ in the visual quality of the edit. EditReward-Compass (2026) holds 2,251 human-verified preference pairs of edited images produced by the same editing models under one instruction, as a reward model meets them during RL; higher is better.
Metric: Preference accuracy (%) on the pairs that differ in visual quality on EditReward-Compass (2,251 human-verified preference pairs of edited images sampled from the same editing models under the same instruction), general MLLMs prompted with the Edit-Compass rubric; thinking disabled unless the row is a thinking run; higher is better. Source: arxiv.org. Saturation forecast: Around January 2028. 18 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 27B | 58.78 | #380 |
| 2 | Qwen 3.6 27B | 57.43 | #358 |
| 3 | Qwen 3.5 35B A3B | 56.08 | #494 |
| 4 | Qwen 3.6 35B A3B | 56.08 | #485 |
| 5 | GPT-4.1 | 53.38 | #506 |
| 6 | Gemma 4 31B | 49.32 | #390 |
| 7 | Qwen 3.5 9B | 48.98 | #838 |
| 8 | Gemini 3 Flash | 48.65 | #180 |
| 9 | Qwen 3.5 2B | 44.66 | #5380 |
| 10 | Gemini 3.1 Pro (Preview) | 44.59 | #102 |
| 11 | Gemma 4 26B A4B | 43.92 | #602 |
| 12 | Qwen 3 VL 32B | 36.49 | #624 |
| 13 | Qwen 3 VL 8B | 34.46 | #927 |
| 14 | Qwen 3 VL 4B | 33.78 | #1163 |
| 15 | Qwen 3 VL 30B A3B | 33.78 | #791 |
No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=editreward-compass-visual-quality · How It Works · Data refreshed daily, snapshot 2026-11-22.