---
license: mit
pretty_name: PhySimCode
language:
- en
size_categories:
- 1K<n<10K
task_categories:
- video-text-to-text
tags:
- physics
- simulation
- video
- code-generation
- multimodal
- benchmark
---

# PhySimCode: A Benchmark and Evaluation Method for Physics Video to Code Generation

[**Website**](https://sourajakundu.github.io/PhySimCode/) · [**Paper**](https://arxiv.org/abs/2610.15722) · [**Code**](https://github.com/SourajaKundu/PhySimCode)

Can a multimodal LLM watch a 2D or 3D physics simulation, work out the involved physics laws and parameter values, and write code from scratch that regenerates it?

This dataset is the **PhySimCode evaluation set**: 2,430 physics simulation videos, each paired with the Python code that rendered it, its true parameter values, and a step-by-step physics explanation (chain-of-thought). It is the set on which all models in the paper are evaluated.

## Overview

- **162 physics phenomena** across 9 physical domains (rigid body, spring / wave, articulated, constrained, pendulum, real-life mixed, rolling, rope / granular, gravitation / fluid)
- **15 samples per phenomenon**, each with randomly sampled physical parameters
- **Two engines**: analytic SciPy simulations (2D) and PyBullet rigid-body scenes (3D)
- Drawn from the full PhySimCode corpus of 160,614 procedurally generated samples

| Engine | Folder | Phenomena | Samples | Video |
|---|---|---|---|---|
| SciPy (2D) | `scipy/` | 87 | 1,305 | 360×180, 10 FPS |
| PyBullet (3D) | `kubric/` | 75 | 1,125 | 240×240, 10 FPS |
| **Total** | | **162** | **2,430** | |

## Structure

```
scipy/<experiment>/<sample_id>/
    video.mp4            # input to the model
    simulation_code.py   # ground-truth code that renders video.mp4
    cot.json             # ground-truth chain-of-thought
    params.json          # sampled parameter values
kubric/<experiment>/<sample_id>/
    ...                  # same four files
```

Every `cot.json` contains `simulation`, `physical_observation`, `physical_law` (name, statement, LaTeX formula), `ode_system`, `derivation_steps`, `numerical_method`, `code_validation`, `expected_behavior` and `parameters` (value, unit and description for every parameter).

- SciPy samples also include `physical_constants_derived` (quantities computed from the parameters).
- PyBullet samples also include `category`, `seed`, `modeling_assumptions`, `impulse_model`, `pybullet_numerical_model` and `visual_scene`.

## Task

Given only `video.mp4` and the engine name (`scipy` or `kubric`), a model must output:

1. a chain-of-thought JSON with `simulation`, `physical_observation`, `physical_law` and `parameters`, and
2. a self-contained Python script that regenerates the video.

Models are scored on physical-law correctness and equivalence (three LLM judges), parameter recovery within ±20%, code compilation / execution / video output, CodeBLEU, and DINOv2 / VideoCLIP similarity between the regenerated and input videos. The prompt is given in Appendix A.4 of the paper, and the evaluation code is on [GitHub](https://github.com/SourajaKundu/PhySimCode).

## Download

```bash
pip install -U huggingface_hub
hf download SourajaKundu123/PhySimCode --repo-type dataset --local-dir benchmarking_data
```

```python
from huggingface_hub import snapshot_download
snapshot_download("SourajaKundu123/PhySimCode", repo_type="dataset", local_dir="benchmarking_data")
```

## Citation

```bibtex
@misc{kundu2026physimcode,
  title         = {PhySimCode: A Benchmark and Evaluation Method for
                   Physics Video to Code Generation},
  author        = {Kundu, Souraja and Gupta, Aditya and Julin, Joel and
                   Zhao, Yizhou and Xie, Liuyue and Jeni, Laszlo A.},
  year          = {2026},
  eprint        = {2610.15722},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  url           = {https://arxiv.org/abs/2610.15722}
}
```

## Contact

Souraja Kundu, [sourajak@cs.cmu.edu](mailto:sourajak@cs.cmu.edu)