SAILResearch/awesome-foundation-model-leaderboards

Covers domain-specific foundation model leaderboards across 15+ modalities (text, code, image, video, agent, medical, etc.) alongside evaluation infrastructure, datasets, and benchmarking tools. Includes a curated taxonomy of both leaderboard platforms (Hugging Face, Kaggle, AIcrowd) and operational tooling (backend templates, scrapers, submission handlers) based on empirical analysis of leaderboard workflows and anti-patterns. Provides integrated search capabilities and focuses exclusively on actively maintained, AI-domain leaderboards to help practitioners benchmark and deploy models.

321 stars.