Spaces:
Running
New look + how to get your model on a board
Welcome 👋 This Space mirrors public standings from BenchHub — an open, multi-modal benchmarking platform where every model is scored on the same data + metric, reproducibly.
Browsing: pick a domain (Vision / Audio / NLP), search, and click any benchmark to see the ranked table — best model first, with a green heatmap per metric.
Getting on a board (free): the Space is read-only, so head to the board’s “Submit” link → runbenchhub.com, sign in (🤗 / GitHub / Google), and run the one-line Python client on your model’s predictions. You’re scored on the exact same eval set as everyone else.
Want a benchmark added? Reply here with the dataset + task and we’ll look at adding it — or spin up your own leaderboard from any HF dataset on BenchHub.
95 boards and counting. Have at it 🚀