README.md

---
license: cc-by-4.0
language:
  - en
pretty_name: Video-MME-Logical
---

# Video-MME-Logical

## A Controlled Diagnostic Benchmark for Video Temporal-Logical Reasoning

**Hohin Kwan\*, Hongyu Li\*, Ray Zhang, Manyuan Zhang, Xianghao Kong, Anyi Rao, Jiahao Xie, and Si Liu**

\* Equal contribution.

[Project Page](https://mrakas.github.io/video-mme-logical/) |
[Paper](https://arxiv.org/abs/2606.27828) |
[Code](https://github.com/Mrakas/video-mme-logical) |
[HF Paper](https://huggingface.co/papers/2606.27828)

![Video-MME-Logical benchmark overview](assets/figure1.png)

## Abstract

Video-MME-Logical is a controlled benchmark for **video temporal-logical reasoning**: the ability to maintain, update, and compose evidence as visual states evolve across frames. It organizes evaluation around five temporal-logical operations and uses programmatic generation to control object states, transitions, temporal dependencies, and logical compositions. The benchmark supports difficulty-controlled final-answer evaluation and intermediate-state diagnostics that verify whether a model recovers the required reasoning trace before producing its final answer.

Experiments with state-of-the-art multimodal large language models reveal a substantial human-model gap, especially as temporal-logical complexity increases. Video-MME-Logical provides a scalable testbed for analyzing and improving this capability.

## Benchmark at a Glance

| Scope | Scale | Notes |
|---|---:|---|
| Full benchmark described in the paper | 503,750 videos | 500,000 training videos and 3,750 test videos |
| Current Hugging Face release | 3,750 test samples | 3,750 videos and 900 referenced images |
| Task taxonomy | 25 task categories | Five temporal-logical operations |
| Difficulty settings | 3 levels | Easy, medium, and hard |
| Intermediate-state subset | 8 task categories | Structured reasoning traces with exact-match verification |

> **Release scope:** this repository currently provides the 3,750-example test set. The 500K training videos described in the paper are not included in the current download.

## Task Taxonomy

| Operation | What it evaluates |
|---|---|
| State Tracking | Maintaining hidden or latent object states across visual transformations |
| Sequential Counting | Accumulating discrete evidence over time |
| Temporal Ordering | Recovering the order of state changes, revealed symbols, or event sequences |
| Dynamic Spatiality | Geometric and motion-based inference |
| Structural Composition | Composing spatial structures across viewpoints, occlusions, and partial observations |

Easy, medium, and hard settings increase the temporal horizon and reasoning complexity while preserving the underlying task definition.

## Controlled Construction

Each task category is implemented as an executable program with temporal transitions, scene configuration, metadata construction, and video rendering. Program-recorded metadata supports question construction, exact answer computation, difficulty control, and intermediate-state supervision.

![Video-MME-Logical construction pipeline](assets/figure2.png)

## Released Files

```text
video_mme_logical.zip
`-- video-mme-logical/
    |-- three_level_testset.json
    |-- videos/                  # 3,750 video files
    `-- images/                  # 900 referenced images
```

Each record in `three_level_testset.json` contains:

| Field | Description |
|---|---|
| `id` | Unique sample identifier |
| `parent_major` | High-level task family |
| `major` | Fine-grained task category |
| `difficulty` | Easy, medium, or hard setting |
| `video_path` | Relative path to the sample video |
| `image_paths` | Optional relative paths to referenced images |
| `question` | Evaluation prompt |
| `answer` | Ground-truth answer |

Relative media paths are resolved from the directory containing `three_level_testset.json`.

## Download

Install the current Hugging Face CLI and download the archive:

```bash
python3 -m pip install -U huggingface_hub
mkdir -p data/hf
hf download marcuskwan/video-mme-logical video_mme_logical.zip \
  --type dataset \
  --local-dir data/hf
unzip data/hf/video_mme_logical.zip -d data
```

After extraction, the test manifest is available at:

```text
data/video-mme-logical/three_level_testset.json
```

## Evaluation

The public GitHub repository includes a minimal Gemini API evaluator and a complete reproduction guide:

- [Evaluation script](https://github.com/Mrakas/video-mme-logical/blob/main/eval.py)
- [Reproduction agent guide](https://github.com/Mrakas/video-mme-logical/blob/main/doc_agent.md)

Run a five-example smoke test before a full evaluation:

```bash
git clone https://github.com/Mrakas/video-mme-logical.git
cd video-mme-logical

python3 -m pip install -U huggingface_hub
mkdir -p data/hf
hf download marcuskwan/video-mme-logical video_mme_logical.zip \
  --type dataset \
  --local-dir data/hf
unzip data/hf/video_mme_logical.zip -d data

uv venv .venv
source .venv/bin/activate
uv pip install google-genai huggingface_hub

export GEMINI_API_KEY=your_key_here
python eval.py \
  --dataset data/video-mme-logical/three_level_testset.json \
  --model gemini-3-pro-preview \
  --limit 5 \
  --output runs/gemini-3-pro-preview_smoke_predictions.jsonl
```

The evaluator writes JSONL predictions and prints the final exact-match accuracy. API credentials are not included in this release.

## Selected Zero-Shot Results

Accuracy (%) on Video-MME-Logical. The table is intentionally compact; see the [paper](https://arxiv.org/abs/2606.27828) or [project leaderboard](https://mrakas.github.io/video-mme-logical/#leaderboard) for full results.

| Model | Overall | Easy | Medium | Hard |
|---|---:|---:|---:|---:|
| Human Level | 95.9 | 98.4 | 95.9 | 93.4 |
| Qwen2.5-VL-72B-Instruct | 12.5 | 15.2 | 13.1 | 9.1 |
| Qwen3-VL-30B-A3B-Think | 10.3 | 16.0 | 8.7 | 6.1 |
| GPT-5.4 | 22.7 | 31.7 | 20.3 | 16.1 |
| **Gemini-3.1 Pro** | **28.6** | **33.1** | **24.1** | **20.6** |

Intermediate-state evaluation is substantially harder: human performance is 96.1%, while GPT-5.4 and Gemini-3.1 Pro reach 17.4% and 10.8%, respectively, on Video-MME-Logical-S.

## Citation

```bibtex
@article{kwan2026video,
  title={Video-MME-Logical: A Controlled Diagnostic Benchmark for Video Temporal-Logical Reasoning},
  author={Kwan, Hohin and Li, Hongyu and Zhang, Ray and Zhang, Manyuan and Kong, Xianghao and Rao, Anyi and Xie, Jiahao and Liu, Si},
  journal={arXiv preprint arXiv:2606.27828},
  year={2026}
}
```

## License

The dataset release in this Hugging Face repository is distributed under the [Creative Commons Attribution 4.0 International License](https://creativecommons.org/licenses/by/4.0/). The evaluation code in the [GitHub repository](https://github.com/Mrakas/video-mme-logical) is distributed under the MIT License.