Hugging Face

Hugging Face is the central hub of the open machine-learning ecosystem, hosting millions of models, datasets, and demos and maintaining the widely used Transformers library. It is where most open-weight models are published and distributed, making it the connective tissue of open AI — a neutral platform used by individual researchers and the largest labs alike.

Founded in 2016, Hugging Face began as a chatbot app before pivoting to build open-source ML tools. Its Transformers library became the de facto standard for working with pretrained models, and the Hugging Face Hub grew into the primary place to share and download them.

The Hub hosts models, datasets, and Spaces (hosted demos), and its download counts are a leading indicator of which open models — Llama, Qwen, DeepSeek, Mistral — are actually being used. The company also offers inference endpoints, an enterprise hub, and libraries spanning training, evaluation, and deployment.

Backed by a broad coalition of investors including major tech and chip companies, Hugging Face has stayed deliberately neutral, benefiting from the growth of open models regardless of which lab produces them. Its role as distribution infrastructure gives it outsized influence over how open AI spreads.