Sign in with Hugging Face to submit a run.

Sign in with Hugging Face

Submit a benchmark run

Upload the leaderboard.json that script.py produced after running your evaluation. It's queued for review — a bench-labs moderator (other than you, if you are one) approves it before it goes live.

Prefer the CLI? python3 script.py --model your/model --publish --hf-token <token> submits the same way, no upload needed.