Active
Hugging Face
Model and dataset repositories, AI libraries, demos, and hosted inference.
- Hub repositories for models, datasets, and Spaces
- Libraries including Transformers and Diffusers
- Inference Providers and dedicated Inference Endpoints
Hugging Face combines a model and dataset hub with development libraries and hosted services. Spaces publishes demo applications; Inference Providers connects applications to partner-hosted models, while Inference Endpoints supplies dedicated managed deployments. Each repository sets its own license, including any restrictions on commercial use.
How pricing works
Costs depend on the service: account plans, storage, Space hardware, inference usage, and dedicated endpoints have different meters. Local model use avoids a hosted inference bill but requires suitable hardware. Check the repository license before selecting either route.
Sources
From the archive
- MiniMax releases H3 video model weights 2026-08-03
- Kimi K2 Thinking adds reasoning between tool calls 2025-11-06
- Llama 4 Scout and Maverick add native image understanding 2025-04-05
- Hugging Face introduces smolagents for tool-using AI applications 2024-12-31
- Llama 3.1 adds a 405B model and longer multilingual context 2024-07-23
- Google releases Gemma 2B and 7B model weights 2024-02-21