JevalKit Leaderboard

JevalKit Leaderboard

An interactive page that shows JevalKit's measured evaluation results as a chart and a filterable table, with model comparison and data download. For people who pick or study models.

Made by Valen-Team on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: This Space is a static web page that shows the current evaluation results from the JevalKit toolkit. You can read scores in a bar chart, filter a table, compare up to three models and download the data. It shows a fixed snapshot, not a live feed.

Who it suits: People who choose between models or study how they score, such as researchers, engineers and students reading evaluation results.

Industry | Who uses it | How they use it | Example
AI research | Evaluation researchers | Compare measured scores across models | Pick three models and view them side by side
Engineering | ML engineers | Narrow models by parameter size | Move the size slider and read the chart
Education | Students | Learn how benchmark results are reported | Open coverage and calibration views
Data work | Analysts | Reuse the published numbers | Download the data from the page

People who want to run their own evaluations can skip it, because this page only displays results.

Where to use it:

• A browser: the Space runs on Hugging Face with nothing to install
• Windows, macOS or Linux: open the downloaded index.html in a web browser

Use case | Who | What they get back
Compare models | Researchers | A bar chart of scores with size sliders and presets
Filter the table | Engineers | The rows that match a benchmark or metric
Check the details | Analysts | Coverage, calibration and downloadable data

How to set it up:

• Python with pip on Windows, macOS or Linux, only to download the files
• A web browser
• Node, only if you use npm start for a local preview

1. Install the Hugging Face hub package.
2. Download the Space's files into a folder.
3. Move into that folder.
4. Open index.html in a web browser.

pip install -U huggingface_hub
hf download Valen-Team/JevalKit-Leaderboard --repo-type space --local-dir JevalKit-Leaderboard
cd JevalKit-Leaderboard

How to use it: Open the Space page, or the local index.html. The chart starts with JevBench accuracy and shows the top ten matching models, and a Show all control opens the rest. The header has controls for English or Chinese and for light or dark. For a local preview you can run the command below.

npm start

Example: You set a parameter-size range with the slider, look at the top models in the chart, then pick three of them to compare. You can download the underlying data from the page.

For newcomers and experts: If you are new, open the Space and read the chart first. It starts with JevBench accuracy, and the table below it lets you filter by benchmark.

For experts, the page is a set of front-end files that run in the browser and need neither a build nor a server. Only full measurements that qualify get a rank, and models never get one aggregate score across different suites. The public-only Decision Index is labeled apart from the full index.

JevalKit exports a results snapshot
The snapshot is stored in this Space
The page draws the chart and table
You filter, compare and download

Good to know:

• License: apache-2
• Cost to run: none named
• Data it reaches: its own results snapshot, and it stores your language and theme choice locally
• Not for: live results or a single score across different suites, since it shows a versioned snapshot

On Hugging Face, 13 November 2026

• Likes: 5
• Built with: static
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/Valen-Team/JevalKit-Leaderboard
• Files: https://huggingface.co/spaces/Valen-Team/JevalKit-Leaderboard/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.