license: cc-by-4.0
language:
- en
pretty_name: ExeWorldBench
task_categories:
- text-generation
- image-text-to-text
tags:
- code-generation
- text-to-code
- image-to-code
- world-modeling
- blender
- benchmark
- physical-reasoning
- physical-intelligence
- visual-to-code
- visual-coding
- vision-language-model
size_categories:
- 1K<n<10K
ExeWorldBench: Benchmarking Code-Based Executable 3D World Modeling
📄 Paper: Coming Soon
🌐 Project Page: https://helen1p.github.io/ExeWorldBench/
Overview
ExeWorldBench is a benchmark for evaluating world understanding and executable realization through code-based 3D world construction. Rather than merely reproducing what a world looks like, ExeWorldBench evaluates whether models can construct an executable world that works.
Usage
1. Harness-Based Evaluation
For harness-based evaluation, the t1/, t2/, and t3/ directories contain the task files and required assets.
Each task is organized in its own directory and can be executed directly using the corresponding agent harness.
Before running a task, replace the dataset path /xxx/ExeWorldBench in the task instructions with your own path to the dataset.
2. API-Based Evaluation
For direct API-based evaluation, use the prompts provided in the api_prompt/ directory.
T1
All T1 instances share the same prompt templates, located in api_prompt/t1/:
prompt.txt: Replace<image>with the corresponding reference image.prompt_image.txt: Replace<image>and<frames>with the corresponding reference image and reference frames, respectively.
The required images and frames can be found in the corresponding task directories under t1/.
T2
All T2 instances share the same prompt templates, located in api_prompt/t2/:
prompt.txt: Replace<image>with the corresponding reference image.prompt_image.txt: Replace<image>and<frames>with the corresponding reference image and motion reference frames, respectively.
The required images and frames can be found in the corresponding task directories under t2/.
T3
Unlike T1 and T2, each T3 instance has its own task-specific prompt.
All individual prompts are provided in api_prompt/t3/, organized by task. Use the corresponding prompt for each instance directly, rather than applying a shared prompt template.
3. Score Evaluation
The prompts for the judge model are provided in the eval/ directory.
Contact
For questions, please contact:
Citation
If you find ExeWorldBench useful in your research, please consider citing our paper: