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

Chinese language startup DeepSeek has launched an up to date model of its R1 reasoning AI mannequin on the developer platform Hugging Face after asserting it in a WeChat message Wednesday morning.

The up to date R1, which is below a permissive MIT license, that means it may be used commercially, is a “minor” improve, in line with DeepSeek’s WeChat announcement. The Hugging Face repository doesn’t comprise an outline of the mannequin — solely configuration information and weights, the inner parts of a mannequin that information its habits.

Weighing in at 685 billion parameters in dimension, the up to date R1 is kind of hefty. (“Parameters” is synonymous with “weights.”) With out modification, the mannequin doubtless can’t run on consumer-grade {hardware}.

DeepSeek rose to prominence earlier this 12 months following the discharge of R1, which gave fashions from OpenAI a run for his or her cash. The startup has raised the ire of some regulators stateside, who argue that DeepSeek’s know-how poses a nationwide safety threat.