# Image editing preference evaluation on EditReward-Bench Source: [Image editing preference evaluation on EditReward-Bench]() Collection index: [llms.txt]() Data updated: 2026-05-12T08:24:17.598Z Evaluation of image editing models based on user preference judgments using the EditReward-Bench dataset, typically measured by accuracy metrics. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Image editing preference evaluation | | Dataset (including any stated split) | EditReward-Bench | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Gemini 3.1 Pro | 63.27 | Not recorded | 6a02b424dbc836b87d8246e4 | | Gemini 3.1 Pro + ARR | 63.27 | Not recorded | 6a02b424dbc836b87d8246e4 | | Gemini 3.1 Pro | 61.23 | Not recorded | 6a02b424dbc836b87d8246e4 | | Gemini 3.1 Pro | 61.23 | Not recorded | 6a02b424dbc836b87d8246e4 | | GPT-5 | 61.01 | Not recorded | 6a02b424dbc836b87d8246e4 | | GPT-5 + ARR | 61.01 | Not recorded | 6a02b424dbc836b87d8246e4 | | GPT-5 | 57.53 | Not recorded | 6a02b424dbc836b87d8246e4 | | GPT-5 | 57.53 | Not recorded | 6a02b424dbc836b87d8246e4 | | Qwen3-VL-8B | 57.22 | Not recorded | 6a02b424dbc836b87d8246e4 | | Qwen3vl-8B + ARR | 57.22 | Not recorded | 6a02b424dbc836b87d8246e4 | | EditReward | 56.45 | Not recorded | 6a02b424dbc836b87d8246e4 | | EditReward | 56.45 | Not recorded | 6a02b424dbc836b87d8246e4 | | Qwen3-VL-8B | 54.01 | Not recorded | 6a02b424dbc836b87d8246e4 | | Qwen3-VL-8B | 54.01 | Not recorded | 6a02b424dbc836b87d8246e4 | ### Gemini 3.1 Pro Accuracy: 63.27. Result ID: 6a02e39f47f34ce924e59ed0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"ARR":"true"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro + ARR Accuracy: 63.27. Result ID: 6a02e39f47f34ce924e59ef8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"ARR (Ours)","Evaluation Protocol":"ARR"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro Accuracy: 61.23. Result ID: 6a02e39f47f34ce924e59ecf. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"ARR":"false"}. This record does not assert independent reproduction. ### Gemini 3.1 Pro Accuracy: 61.23. Result ID: 6a02e39f47f34ce924e59ef5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"VLM-as-Judge (Direct)","Evaluation Protocol":"Direct"}. This record does not assert independent reproduction. ### GPT-5 Accuracy: 61.01. Result ID: 6a02e39f47f34ce924e59ece. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"ARR":"true"}. This record does not assert independent reproduction. ### GPT-5 + ARR Accuracy: 61.01. Result ID: 6a02e39f47f34ce924e59ef7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"ARR (Ours)","Evaluation Protocol":"ARR"}. This record does not assert independent reproduction. ### GPT-5 Accuracy: 57.53. Result ID: 6a02e39f47f34ce924e59ecd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"ARR":"false"}. This record does not assert independent reproduction. ### GPT-5 Accuracy: 57.53. Result ID: 6a02e39f47f34ce924e59ef4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"VLM-as-Judge (Direct)","Evaluation Protocol":"Direct"}. This record does not assert independent reproduction. ### Qwen3-VL-8B Accuracy: 57.22. Result ID: 6a02e39f47f34ce924e59ecc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"ARR":"true"}. This record does not assert independent reproduction. ### Qwen3vl-8B + ARR Accuracy: 57.22. Result ID: 6a02e39f47f34ce924e59ef6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"ARR (Ours)","Evaluation Protocol":"ARR"}. This record does not assert independent reproduction. ### EditReward Accuracy: 56.45. Result ID: 6a02e39f47f34ce924e59eca. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### EditReward Accuracy: 56.45. Result ID: 6a02e39f47f34ce924e59ef2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category":"Trained Reward Model"}. This record does not assert independent reproduction. ### Qwen3-VL-8B Accuracy: 54.01. Result ID: 6a02e39f47f34ce924e59ecb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"ARR":"false"}. This record does not assert independent reproduction. ### Qwen3-VL-8B Accuracy: 54.01. Result ID: 6a02e39f47f34ce924e59ef3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"VLM-as-Judge (Direct)","Evaluation Protocol":"Direct"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.