| --- |
| license: cc-by-nc-4.0 |
| pretty_name: VisualWorldBench |
| task_categories: |
| - visual-question-answering |
| - image-classification |
| - object-detection |
| language: |
| - en |
| size_categories: |
| - 10K<n<100K |
| --- |
| |
| <h1 align="center">VisualWorldBench</h1> |
|
|
| <h3 align="center">A Fine-Grained Multi-Task Benchmark for Evaluating<br>Visual World Knowledge in MLLMs</h3> |
|
|
| <p align="center"> |
| Chen Chen<sup>1,2</sup>  |
| Zheming Liang<sup>1,2</sup>  |
| Jiaqi Wang<sup>2,3,*</sup> |
| </p> |
| |
| <p align="center"> |
| <sup>1</sup> University of Science and Technology of China<br> |
| <sup>2</sup> Shanghai Innovation Institute  <sup>3</sup> JD.com<br> |
| <sub>* Corresponding author</sub> |
| </p> |
|
|
| <p align="center"> |
| <a href="#overview">Overview</a> ·  |
| <a href="#download">Download</a> ·  |
| <a href="#data-format">Data Format</a> ·  |
| <a href="#citation">Citation</a> |
| </p> |
|
|
| ## Overview |
|
|
| VisualWorldBench evaluates fine-grained visual world knowledge in multimodal large language models. It contains **32,541 evaluation instances** across four tasks and **15,479 ground-truth categories**, spanning eight visual domains. |
|
|
| | Task | Annotation file | Instances | Metric | |
| |---|---|---:|---| |
| | MCQ | `mcq.jsonl` | 14,717 | Accuracy | |
| | Grid-Sel (2×2) | `sel_2x2.jsonl` | 4,085 | Accuracy | |
| | Grid-Sel (3×3) | `sel_3x3.jsonl` | 4,125 | Accuracy | |
| | Grid-Loc (2×2) | `loc_2x2.jsonl` | 3,959 | IoU@0.5 | |
| | Grid-Loc (3×3) | `loc_3x3.jsonl` | 3,446 | IoU@0.5 | |
| | Open-QA | `open_qa.jsonl` | 2,209 | EM + LLM Judge | |
|
|
| The overall score is the mean of MCQ, Grid-Sel (averaged over both grid sizes), Grid-Loc (averaged over both grid sizes), and corrected Open-QA accuracy. VisualWorldBench-50K is a separate training resource. |
|
|
| ## Download |
|
|
| ```bash |
| hf download orange0318/VisualWorldBench --repo-type dataset --local-dir ./VisualWorldBench |
| cd VisualWorldBench |
| ``` |
|
|
| Grid images are included. The **16,095 unique single images** used by MCQ and Open-QA are provided through source links. Download direct image URLs to the local paths specified by the annotations. For images distributed through upstream datasets, obtain and extract those datasets separately, then locate the images using the recorded source paths. |
|
|
| ## Data Format |
|
|
| Annotations are stored as JSONL, with one evaluation instance per line. Each record has a `qid`, an `image` path, and task-specific fields: |
|
|
| | Task | Main fields | |
| |---|---| |
| | MCQ | `options`, `gt_index` | |
| | Grid-Sel | `target_class_name`, `target_cell`, `answer` | |
| | Grid-Loc | `target_class_name`, `bbox_norm`, `answer` | |
| | Open-QA | `question`, `gt_aliases_base`, `gt_aliases_expanded`, `gt_ancestors` | |
|
|
| The evaluation code provides prompt templates for Grid-Sel and Grid-Loc. |
|
|
| `bbox_norm` uses `[x1, y1, x2, y2]` coordinates normalized to [0, 1000]. Record metadata includes the visual domain and taxonomy depth. The `meta.domain` label follows the ground-truth category in the taxonomy: `animal`, `plant`, `fungus`, `artifact`, `transport`, `food`, `scene`, or `other`. |
|
|
| Single-image records also contain `image_url` and `image_url_type`. The `image` field specifies where to save the retrieved image locally. Records requiring an upstream dataset additionally include its identifier, relative path and acquisition filename. |
|
|
| ## Citation |
|
|
| ```bibtex |
| @misc{chen2026visualworldbench, |
| title = {VisualWorldBench: A Fine-Grained Multi-Task Benchmark for Evaluating Visual World Knowledge in {MLLMs}}, |
| author = {Chen, Chen and Liang, Zheming and Wang, Jiaqi}, |
| year = {2026} |
| } |
| ``` |
|
|
| ## License |
|
|
| Original benchmark annotations are licensed under [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/). Source images retain their respective copyrights, licenses, and access conditions. |
|
|
| If you encounter any issues downloading the images, please contact the authors. |
|
|
