China’s DeepSeek updates its R1 reasoning AI model

The Chinese artificial intelligence (AI) startup DeepSeek has released an updated version of its AI chatbot that took the world by storm when it was released earlier this year, R1 reasoning model.

The AI model is now available on the developer platform Hugging Face, as per the company’s official WeChat account.

According to DeepSeek, this update is a “minor” enhancement to the original R1 model. The new version is distributed under the permissive MIT license, allowing developers and businesses to use it freely for commercial purposes.

The updated model features a massive 685 billion parameters, placing it among the largest publicly released AI models to date.

For those unversed, parameters, also known as weights, are the internal elements that dictate how an AI model functions.

Due to its enormous size, the model is unlikely to be compatible with standard consumer hardware without further optimization.

While the Hugging Face repository currently lacks a detailed description of the model, it does include the essential configuration files and weight data needed for developers to begin working with it.

Despite gaining significant attention earlier this year, the startup has also drawn scrutiny from US regulators, with some officials raising concerns about the potential national security risks posed by its technology.