AI Phone Leaderboard

AI Phone Leaderboard

Shows and analyzes AI benchmark results from mobile devices in a Streamlit dashboard with charts and filters. For people comparing how phones run AI models.

Made by a-ghorbani 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: AI Phone Leaderboard is a Streamlit dashboard for viewing and analyzing AI benchmark results from mobile devices. It reads its data from a Firebase Firestore database and offers interactive charts and filters. You can open it in a browser without installing anything.

Who it suits: It suits mobile developers, researchers and device reviewers who want to compare how different phones run AI workloads, and teams that want to host their own benchmark dashboard.

Industry | Who uses it | How they use it | Example
Mobile development | App developers | Check how devices handle on-device AI | Filter the results by device
AI research | Researchers | Analyze benchmark submissions across phones | Compare runs with the interactive charts
Hardware reviews | Reviewers | Compare phones on AI benchmark results | Sort devices by a result

If you do not care about mobile AI performance, you can skip it.

Where to use it:

• A web browser, on the Space's page on Hugging Face
• A local copy of the files, run with Python under Windows, macOS or Linux
• Your own Hugging Face Space, set up with the Streamlit SDK

Use case | Who | What they get back
Browse benchmark results | Anyone | Interactive visualizations of phone results
Filter by device | Researchers | A narrowed view of the results
Host a dashboard | Developers | Their own copy connected to their own Firestore data

How to set it up:

• A web browser, which is all the online page needs
• Python 3.10 or newer, for a local copy
• A Firebase project with a Firestore database, for your own copy
• A read-only service account for the dashboard, with its credentials kept out of git
• A Hugging Face account, to deploy your own Space

1. Open the Space's page on Hugging Face to use it online. Nothing needs installing.
2. For a local copy, install the Hugging Face hub package and download the Space's files.
3. Move into the downloaded folder.
4. Install Streamlit 1.41.0, then the packages from requirements.txt.
5. Start the app with streamlit run main.py.

pip install -U huggingface_hub
hf download a-ghorbani/ai-phone-leaderboard --repo-type space --local-dir ai-phone-leaderboard
cd ai-phone-leaderboard
pip install streamlit==1.41.0
pip install -r requirements.txt
streamlit run main.py

The README offers a make-based route for development.

make venv
make setup-dev
make run

For your own copy, store the Firebase credentials as a secret named FIREBASE_CREDENTIALS, in a local secrets file or in the Space's repository secrets.

How to use it: Open the dashboard, look at the visualizations and use the filters to narrow the results. The project's make commands cover day-to-day development.

Tool | What it does
make help | Shows the available commands
make setup-dev | Sets up the development environment
make setup-prod | Sets up the production environment
make run | Runs the Streamlit application
make lint | Runs the code linter
make format | Formats the code with black
make test | Runs the tests
make clean | Clears cache files

Example: Open the dashboard and filter the results to one device. The charts update to show only the benchmark runs that match.

For newcomers and experts: Open the Space page and play with the filters first. Setting up Firebase is only needed if you want a dashboard of your own.

The app expects a Firestore collection laid out as benchmarks, then v1, then submissions, with one document per benchmark. Each document holds benchmarkResult, deviceInfo and metadata. The README recommends a dedicated service account with minimal read-only permissions, and secrets that never go into git.

Phones submit benchmark results
Results are stored in Firebase Firestore
The dashboard reads them with a read-only account
Streamlit shows charts and filters

Good to know: The dashboard only reads data, and the credentials must stay out of the repository.

• License: MIT
• Cost to run: a Firebase project with Firestore and a read-only service account, for your own copy
• Data it reaches: benchmark submissions in Firestore, including device information, read only

On Hugging Face, 13 November 2026

• Likes: 127
• Built with: streamlit
• License: mit

Links

• Space: https://huggingface.co/spaces/a-ghorbani/ai-phone-leaderboard
• Files: https://huggingface.co/spaces/a-ghorbani/ai-phone-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.