EXPLORINEER
en

Archive — US — techcrunch

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

Chinese startup DeepSeek has released an updated R1 reasoning AI model on Hugging Face, described as a 'minor' upgrade with 685 billion parameters. The model is commercially usable under an MIT license. This release has raised concerns among regulators regarding national security risks associated with its technology.

Key points

  • DeepSeek updates its R1 reasoning AI model and releases it on Hugging Face.
  • The update is termed a 'minor' upgrade by DeepSeek in a WeChat announcement.
  • The R1 model has 685 billion parameters, making it heavy and likely unsuitable for consumer-grade hardware.
  • The model is available under an MIT license, allowing commercial use.
  • DeepSeek has been in the spotlight for challenging established AI models from OpenAI.
  • Regulatory concerns have emerged due to perceived national security risks posed by DeepSeek's technology.

Why it matters

  • AI development is increasingly scrutinized under national security frameworks, particularly for startups like DeepSeek.
  • The rise of reasoning AI models reflects a trend in 2025 where AI must be robust yet ethically managed to prevent misuse.
  • There is a growing conversation about the impacts of advanced AI technologies on consumer safety and national security.

DeepSeek's updated R1 AI model signifies advancements in reasoning technology but raises critical concerns about its implications for security and regulation in the AI landscape.