Multi-way preference ranking on EDITREWARD-BENCH
66.2Preference Score (K=2)REWARDHARNESS (Gemini-2.0-Flash)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| REWARDHARNESS (Gemini-2.0-Flash)Sub-Agent=Gemini-2.0-Flash2026.05 | 66.2 | 45.3 | 13.5 | — | |
| Gemini-2.5-FlashMethod Category=Proprietary Models2025.09 | 58.61 | 39.86 | 12.16 | 38.02 | |
| Gemini-2.5-FlashModel Type=Proprietary Models2026.05 | 58.6 | 39.9 | 12.2 | — | |
| Claude-Haiku-4.5Model Type=Proprietary Models2026.05 | 57.9 | 30.7 | 7.4 | — | |
| REWARDHARNESS (Qwen)Sub-Agent=Qwen2026.05 | 57.9 | 46.7 | 10.8 | — | |
| GPT-5Method Category=Proprietary Models2025.09 | 57.53 | 38.51 | 12.84 | 37.81 | |
| GPT-5Model Type=Proprietary Models2026.05 | 57.5 | 38.5 | 12.8 | — | |
| EditReward (Qwen)Model Type=Open-Source Models, Backbone=Qwen2026.05 | 57 | 36 | 10.8 | — | |
| EDITREWARDBackbone=Qwen2.5-VL-7B2025.09 | 56.99 | 36 | 10.81 | 36.78 | |
| EditReward (MiMo)Model Type=Open-Source Models, Backbone=MiMo2026.05 | 56.5 | 42.7 | 11.5 | — | |
| EDITREWARDBackbone=MiMo-VL-7B-SFT2025.09 | 56.45 | 42.67 | 11.49 | 38.42 | |
| Qwen2.5-VL-7BModel Type=Open-Source Models2026.05 | 52.7 | 24.7 | 3.4 | — | |
| Qwen2.5-VL-7B-InstMethod Category=Open-Source VLMs2025.09 | 52.69 | 24.67 | 3.38 | 29.75 | |
| Gemini-2.0-FlashMethod Category=Proprietary Models2025.09 | 52.43 | 33.33 | 13.51 | 33.47 | |
| Gemini-2.0-FlashModel Type=Proprietary Models2026.05 | 52.4 | 33.3 | 13.5 | — | |
| Qwen2.5-VL-3B-InstMethod Category=Open-Source VLMs2025.09 | 51.07 | 20.27 | 2.71 | 26.86 | |
| Qwen2.5-VL-32B-InstMethod Category=Open-Source VLMs2025.09 | 50.54 | 25.27 | 4.05 | 28.72 | |
| Qwen2.5-VL-32BModel Type=Open-Source Models2026.05 | 50.5 | 25.3 | 4.1 | — | |
| MiMo-VL-7BModel Type=Open-Source Models2026.05 | 49.5 | 30.4 | 9.5 | — | |
| MiMo-VL-7B-SFT-2508Method Category=Open-Source VLMs2025.09 | 49.46 | 30.41 | 9.46 | 31.19 | |
| GPT-4oModel Type=Proprietary Models2026.05 | 45.7 | 27.3 | 7.3 | — | |
| GPT-4oMethod Category=Proprietary Models2025.09 | 45.69 | 27.33 | 7.31 | 28.31 | |
| RandomMethod Category=Baseline2025.09 | 25.81 | 11.33 | 1.35 | 13.84 |