# Multimodal Preference Evaluation on MMRB2 Source: [Multimodal Preference Evaluation on MMRB2]() Collection index: [llms.txt]() Data updated: 2026-03-10T11:30:37.237Z Evaluation of model performance on the MMRB2 dataset using Multimodal Preference Evaluation, focusing on T2I, Editing, Interleaved, and Reasoning accuracy. 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 | Multimodal Preference Evaluation | | Dataset (including any stated split) | MMRB2 | | Metric | T2I Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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 - [T2I Accuracy](): Higher is better; unit: PERCENT. - [Editing Accuracy](): Higher is better; unit: PERCENT. - [Interleaved Accuracy](): Higher is better; unit: PERCENT. - [Reasoning Accuracy](): Higher is better; unit: PERCENT. - [Average Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | T2I Accuracy | Result date | Report paper | | --- | --- | --- | --- | | MJ1 | 80.2 | Not recorded | 69afa5952d603e6de3e43bf3 | | Gemini 3 Pro | 74.4 | Not recorded | 69afa5952d603e6de3e43bf3 | | GPT-5 | 70.5 | Not recorded | 69afa5952d603e6de3e43bf3 | | Gemini 2.5 Pro | 70.5 | Not recorded | 69afa5952d603e6de3e43bf3 | | GPT-4.1 | 65.8 | Not recorded | 69afa5952d603e6de3e43bf3 | | Qwen3-VL-32B | 64.1 | Not recorded | 69afa5952d603e6de3e43bf3 | | Gemini 2.5 Flash | 63.1 | Not recorded | 69afa5952d603e6de3e43bf3 | | Qwen3-VL-235B-A22B | 62 | Not recorded | 69afa5952d603e6de3e43bf3 | | GPT-4o | 60.3 | Not recorded | 69afa5952d603e6de3e43bf3 | | Qwen3-VL-30B-A3B | 60 | Not recorded | 69afa5952d603e6de3e43bf3 | | Qwen3-VL-8B | 59.4 | Not recorded | 69afa5952d603e6de3e43bf3 | | Qwen2.5-VL-72B | 59.1 | Not recorded | 69afa5952d603e6de3e43bf3 | | Gemma 3 27B | 58.3 | Not recorded | 69afa5952d603e6de3e43bf3 | | Gemma 3 12B | 56 | Not recorded | 69afa5952d603e6de3e43bf3 | | Gemma 3 4B | 51.7 | Not recorded | 69afa5952d603e6de3e43bf3 | | Qwen2.5-VL-7B | 50.4 | Not recorded | 69afa5952d603e6de3e43bf3 | ### MJ1 T2I Accuracy: 80.2. Result ID: 69b000cd6dbb7075b9597621. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family/Base Model":"Qwen3-VL-30B-A3B","Fine-tuning Protocol (e.g., LoRA)":"LoRA","Active Parameters":"3B"}. This record does not assert independent reproduction. ### Gemini 3 Pro T2I Accuracy: 74.4. Result ID: 69b000cd6dbb7075b9597620. [Reporting paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"API-based"}. This record does not assert independent reproduction. ### GPT-5 T2I Accuracy: 70.5. Result ID: 69b000cd6dbb7075b959761d. [Reporting paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"API-based"}. This record does not assert independent reproduction. ### Gemini 2.5 Pro T2I Accuracy: 70.5. Result ID: 69b000cd6dbb7075b959761f. [Reporting paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"API-based"}. This record does not assert independent reproduction. ### GPT-4.1 T2I Accuracy: 65.8. Result ID: 69b000cd6dbb7075b959761c. [Reporting paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"API-based"}. This record does not assert independent reproduction. ### Qwen3-VL-32B T2I Accuracy: 64.1. Result ID: 69b000cd6dbb7075b9597618. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"32B"}. This record does not assert independent reproduction. ### Gemini 2.5 Flash T2I Accuracy: 63.1. Result ID: 69b000cd6dbb7075b959761e. [Reporting paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"API-based"}. This record does not assert independent reproduction. ### Qwen3-VL-235B-A22B T2I Accuracy: 62. Result ID: 69b000cd6dbb7075b959761a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"22B"}. This record does not assert independent reproduction. ### GPT-4o T2I Accuracy: 60.3. Result ID: 69b000cd6dbb7075b959761b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"API-based"}. This record does not assert independent reproduction. ### Qwen3-VL-30B-A3B T2I Accuracy: 60. Result ID: 69b000cd6dbb7075b9597619. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"3B"}. This record does not assert independent reproduction. ### Qwen3-VL-8B T2I Accuracy: 59.4. Result ID: 69b000cd6dbb7075b9597617. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"8B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-72B T2I Accuracy: 59.1. Result ID: 69b000cd6dbb7075b9597616. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"72B"}. This record does not assert independent reproduction. ### Gemma 3 27B T2I Accuracy: 58.3. Result ID: 69b000cd6dbb7075b9597614. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"27B"}. This record does not assert independent reproduction. ### Gemma 3 12B T2I Accuracy: 56. Result ID: 69b000cd6dbb7075b9597613. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"12B"}. This record does not assert independent reproduction. ### Gemma 3 4B T2I Accuracy: 51.7. Result ID: 69b000cd6dbb7075b9597612. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"4B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B T2I Accuracy: 50.4. Result ID: 69b000cd6dbb7075b9597615. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Category (Open-source vs API-based)":"Open-source","Active Parameters":"7B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.