Hugging Face

Hugging Face operates a central platform for developing and sharing machine learning models and is considered one of the most important hubs for the open-source AI community. Its Model Hub hosts more than two million freely accessible models, complemented by over 500,000 datasets for a wide range of ML tasks.

At the core of the ecosystem is the Transformers library for PyTorch, with more than 166,000 GitHub stars, widely regarded as the de facto standard for deploying modern language and vision models. With "Spaces," Hugging Face also offers a way to host and try out AI applications directly — more than one million applications have already been published this way. Other open-source tools such as Diffusers, Safetensors, TRL, and PEFT round out the ecosystem with specialized training and optimization methods.

Commercially, Hugging Face offers managed Inference Endpoints, access to more than 45,000 models via Inference Providers, and team/enterprise plans starting at $20 per user per month, plus GPU compute starting at $0.60 per hour.

The company says more than 50,000 organizations use the platform, including Meta, Google, Amazon, and Microsoft — a sign of how central Hugging Face has become to practical AI deployment, well beyond individual research projects.

The company's name comes from the yellow emoji with two open hands, which has served as its logo since founding — a deliberately unconventional look for a company that, at its core, provides highly technical infrastructure for machine learning.