TREE NEWS reports: Alibaba, Tencent and Moonshot AI have reduced the offensive cyber capabilities of their newly released models while strengthening defensive capabilities. Alibaba’s latest flagship Qwen series model does not publicly release its model weights, preventing users from downloading and customizing the full version.
Alibaba, Tencent and Moonshot AI Limit Offensive Cyber Capabilities in Latest Models
The closed-weight approach on Alibaba's flagship Qwen model marks a notable shift for a company whose open releases helped build its following, and pairing that with curbed offensive cyber capability suggests safety and misuse concerns are now shaping release decisions at major Chinese labs. This affects developers who relied on downloadable weights for customization and red-teaming, and it widens the gap between open and gated model access in the ecosystem. Whether the defensive-strengthening claims hold up under independent testing, and whether other labs follow the closed-weight path, is the open question worth watching.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.