United States Commerce Department said in a statement on Thursday that the US Center for AI Standards and Innovation CAISI and the United Kingdom Artificial Intelligence Security Institute's (UK AISI) preliminary analysis of Kimi K3 showed that the AI model built by China's Moonshot AI "performed significantly below the leading US frontier AI models."
In the evaluation, which focused on Kimi K3's cyber capabilities, it was mentioned that the Chinese AI model specifically performs significantly worse compared to the US models "when tasked to develop exploits [and] to attack a simulated corporate network." Meanwhile, the report noted that Kimi K3's performance is above GLM-5.2, while the model's safeguards "allow assistance with agentic cyber exploit development."
https://breakingthenews.net/Article/US:-Analysis-shows-Kimi-K3-behind-our-AI-models/66763611
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.