The independent, open leaderboard for AI models.
We score open-weight AI models with transparent, reproducible signals instead of hype: benchmarks, efficiency, community, recency, and reproducibility, plus head-to-head ELO and free embeddable badges.
HuggingFace tells you what models exist. ModelRank tells you which ones are actually good, using a weighted score and head-to-head win probabilities rather than a black box. We do not train or host models, so we have no reason to rank our own higher.
Python · FastAPI · Gradio · LangChain · LlamaIndex
⭐ Star the repository to help an independent project grow.