DeepSeek updates its R1 reasoning AI model, releases it on Hugging Face

Chinese startup DeepSeek has launched an updated version of its R1 reasoning AI model on the Hugging Face platform, following an announcement on WeChat. The update is described as a "minor" upgrade and is available under the permissive MIT license, allowing for commercial use. The model, which is substantial at 685 billion parameters, is unlikely to run on consumer-grade hardware without modifications. The Hugging Face repository includes only configuration files and weights, lacking a detailed description of the model.

DeepSeek gained attention earlier this year with the initial release of R1, which competed closely with models from OpenAI. However, the company has faced criticism from some U.S. regulators who view its technology as a potential national security threat.

Key takeaways

Chinese startup DeepSeek released an updated version of its R1 reasoning AI model on Hugging Face.
The updated R1 model is under a permissive MIT license, allowing commercial use.
With 685 billion parameters, the updated R1 model is too large for consumer-grade hardware.
DeepSeek's technology has raised national security concerns among some U.S. regulators.

DeepSeek updates its R1 reasoning AI model, releases it on Hugging Face | TechCrunch

Key takeaways

Discussion (0)