Skip to main content

Ali's Zhenwu M890 Super Node Now Runs Qwen3.8, Available on BaiLian

Alibaba Cloud has made a significant leap in AI infrastructure. On July 23, the company announced that its Zhenwu M890 super node is now fully compatible with Qwen3.8, its latest flagship large language model. This integration is already live on the Alibaba Cloud BaiLian platform, providing inference services to users.

What makes this announcement stand out is the sheer scale of the model. Qwen3.8 packs a whopping 2.4 trillion parameters. To put that in perspective, running a model of this size is no small feat. Traditional AI clusters often struggle with communication bottlenecks between nodes, leading to performance issues when handling high-throughput, low-latency tasks. Typically, you'd need to split the model across dozens or even hundreds of high-speed GPU cards just to keep things running smoothly.

The Zhenwu M890 super node tackles this head-on. It features optimized high-speed interconnect technology and a revamped communication architecture. These improvements overcome the transmission limitations that plague conventional clusters, ensuring that Qwen3.8 runs both stably and quickly during inference. In fact, Alibaba Cloud claims this is the first super node system in China to successfully support a large model with over 2 trillion parameters—a notable milestone for domestic software and hardware collaboration.

For enterprises and developers, this means direct access to efficient trillion-parameter inference capabilities. No more wrestling with complex distributed setups. The service is available now on the BaiLian platform, laying the groundwork for next-generation ultra-large-scale AI applications. Whether it's for advanced natural language processing, complex reasoning, or other AI-driven tasks, this setup promises to deliver the computational muscle needed.

Key Points

  • Model Scale: Qwen3.8 has 2.4 trillion parameters, requiring robust infrastructure.
  • Breakthrough: Zhenwu M890 is China's first super node to support such a large model.
  • Technology: Optimized high-speed interconnect and communication architecture eliminate traditional bottlenecks.
  • Availability: Inference services are live on Alibaba Cloud's BaiLian platform.
  • Impact: Enables commercial deployment of next-gen ultra-large AI applications.