| <div class="section" style="padding: 10px 0;"> | |
| <p class="muted"> | |
| The <span class="emphasis">3D Arena</span> leaderboard evaluates leading generative 3D | |
| models. It is inspired by | |
| <a class="emphasis" target="_blank" href="https://huggingface.co/spaces/TTS-AGI/TTS-Arena" | |
| >TTS-Arena</a | |
| > | |
| and | |
| <a class="emphasis" target="_blank" href="https://chat.lmsys.org/">Chatbot Arena</a>. | |
| </p> | |
| <p class="muted"> | |
| Read more about the platform and methodology in the official paper: | |
| <a class="emphasis" target="_blank" href="https://huggingface.co/papers/2506.18787" | |
| >3D Arena: An Open Platform for Generative 3D Evaluation</a | |
| >. | |
| </p> | |
| </div> | |
| <div class="section"> | |
| <h3 class="emphasis">Motivation</h3> | |
| <p class="muted"> | |
| Evaluating generated 3D quality quantitatively is challenging, and there is no standard | |
| practice for reporting critical factors like mesh topology in real-world applications. | |
| </p> | |
| <p class="muted"> | |
| This leaderboard displays usable, final results side-by-side for the community to rank. | |
| </p> | |
| </div> | |
| <div class="section"> | |
| <h3 class="emphasis">Criteria</h3> | |
| <p class="muted"> | |
| Models must handle <span class="emphasis">image</span> (.png, .jpg) inputs and produce | |
| <span class="emphasis">mesh</span> | |
| (.obj, .glb) or <span class="emphasis">splat</span> (.splat, .ply) outputs. They should run end-to-end | |
| without human intervention, including UV unwrapping, texture mapping, and other post-processing. | |
| </p> | |
| <p class="muted"> | |
| To preview how your results will look in the leaderboard, upload them to any <a | |
| class="emphasis" | |
| target="_blank" | |
| href="https://huggingface.co/spaces/gradio/model3D">gradio/model3d</a | |
| > viewer. | |
| </p> | |
| </div> | |
| <div class="section"> | |
| <h3 class="emphasis">Contributing</h3> | |
| <p class="muted"> | |
| All inputs and outputs are publicly available <a | |
| class="emphasis" | |
| target="_blank" | |
| href="https://huggingface.co/datasets/dylanebert/3d-arena">here</a | |
| >. Input image URLs are provided | |
| <a | |
| class="emphasis" | |
| target="_blank" | |
| href="https://huggingface.co/datasets/dylanebert/3d-arena/raw/main/inputs.txt">here</a | |
| >. | |
| </p> | |
| <p class="muted"> | |
| To contribute a model, run your model on all input images and create a pull request with | |
| your model's output. Refer to other examples in the dataset to understand the expected | |
| format. | |
| </p> | |
| </div> | |
| <div class="section"> | |
| <h3 class="emphasis">Citing</h3> | |
| <div class="pre-wrapper"> | |
| <pre> | |
| <code> | |
| @misc{ebert20253darenaopenplatform, | |
| title = {3D Arena: An Open Platform for Generative 3D Evaluation}, | |
| author = {Dylan Ebert}, | |
| year = {2025}, | |
| eprint = {2506.18787}, | |
| archivePrefix = {arXiv}, | |
| primaryClass = {cs.CV}, | |
| url = {https://arxiv.org/abs/2506.18787}, | |
| } | |
| </code> | |
| </pre> | |
| </div> | |
| </div> | |