README.md

---
license: mit
tags:
  - benchmark
linked_spaces:
  - FlyaiLab/ecommerce_last_exam_leaderboard
dataset_info:
  features:
    - name: instance_id
      dtype: string
    - name: domain
      dtype: string
    - name: config
      dtype: string
    - name: user_question
      dtype: string
    - name: docker_image
      dtype: string
  splits:
    - name: test
      num_examples: 120
configs:
  - config_name: default
    data_files:
      - split: test
        path: data/test-*
  - config_name: travel
    data_files:
      - split: test
        path: data/travel/test-*
  - config_name: e_commerce
    data_files:
      - split: test
        path: data/e_commerce/test-*
---

# E-Commerce Last Exam

<iframe
  src="https://FlyaiLab-ecommerce_last_exam_leaderboard.hf.space"
  frameborder="0"
  width="100%"
  height="600"
></iframe>

A benchmark for evaluating LLM agents on **120 real-world travel planning and e-commerce tool-use tasks**. Each task runs in an isolated Docker container with domain-specific CLI tools and SQLite databases. Agents must search, analyze, and produce structured recommendations.

- **Repository**: [alibaba-flyai/ecommerce_last_exam](https://github.com/alibaba-flyai/ecommerce_last_exam)
- **Evaluation CLI**: [flyai-bench](https://pypi.org/project/flyai-bench/) (`pip install flyai-bench`)
- **Leaderboard**: [FlyaiLab/ecommerce_last_exam_leaderboard](https://huggingface.co/spaces/FlyaiLab/ecommerce_last_exam_leaderboard)

## Dataset Summary

E-Commerce Last Exam consists of 120 tasks across two configs:

| Config | Tasks | Domains |
|--------|-------|---------|
| `travel` | 77 | Hotel booking, transport routing, attraction planning |
| `e_commerce` | 43 | Travel gear, food, electronics, lifestyle shopping |

Each task provides a natural-language user question and a pre-built Docker image containing the environment (CLI tools, databases, test harness). Agents interact with the environment via tool calls and produce a structured answer scored 0.00 - 1.00.

## Data Fields

| Field | Type | Description |
|-------|------|-------------|
| `instance_id` | string | Unique task identifier (e.g., `attraction_auckland_extreme_sports_415`) |
| `domain` | string | Task domain (e.g., `attraction`, `hotel`, `transport`, `consume`) |
| `config` | string | Benchmark config: `travel` or `e_commerce` |
| `user_question` | string | Natural-language task description |
| `docker_image` | string | Docker image for the task environment |

## Usage

```python
from datasets import load_dataset

# Load all 120 tasks
ds = load_dataset("FlyaiLab/ecommerce_last_exam", split="test")

# Load only travel tasks (77)
ds_travel = load_dataset("FlyaiLab/ecommerce_last_exam", "travel", split="test")

# Load only e-commerce tasks (43)
ds_ecom = load_dataset("FlyaiLab/ecommerce_last_exam", "e_commerce", split="test")
```

## Evaluation

```bash
pip install flyai-bench

# Run evaluation on travel config
flyai-bench run --dataset-config travel --limit 5 --dry-run
flyai-bench run --dataset-config travel

# Generate report and submit
flyai-bench report
flyai-bench submit --model your-model --provider your-provider
```

See the [GitHub repository](https://github.com/alibaba-flyai/ecommerce_last_exam) for full documentation.

## Citation

If you use this benchmark, please cite:

```bibtex
@misc{ecommerce_last_exam,
  title={E-Commerce Last Exam: A Benchmark for LLM Agent Evaluation on Real-World Tool-Use Tasks},
  author={FlyaiLab},
  year={2026},
  url={https://huggingface.co/datasets/FlyaiLab/ecommerce_last_exam}
}
```