Home
Ali Zhenwu M890 Ultra Node Now Compatible with Qwen3.8, Available on BaiLian Platform for Inference

Alibaba Cloud confirmed on July 23 that its Zhenwu M890 super node has successfully integrated with the latest flagship large model, Qwen3.8, making it available for model inference services on the Alibaba Cloud BaiLian platform. This milestone establishes the Zhenwu M890 as China’s first super node system to support the stable operation of a large model exceeding 2 trillion parameters, marking a significant breakthrough in domestic software-hardware collaboration for large model computing scheduling.
Qwen3.8, Alibaba’s newest ultra-large-scale flagship model, features up to 2.4 trillion parameters. Traditional AI clusters often face performance bottlenecks in handling high-throughput, low-latency, and high-concurrency inference demands due to limited inter-node communication bandwidth. To ensure stable and efficient operation at this scale, model parameters typically need to be split and distributed across dozens or even hundreds of high-speed connected GPU cards.
To tackle the distributed computing challenges posed by trillion-parameter models, the Alibaba Zhenwu M890 super node has optimized cluster high-speed interconnection technology and communication architecture, effectively overcoming the transmission limitations of traditional AI clusters. This ensures that large models run “stably and quickly” during inference. With this service now available on the Alibaba Cloud BaiLian platform, enterprises and developers can directly access efficient trillion-parameter inference capabilities, laying the computing foundation for the commercial deployment of next-generation ultra-large-scale AI applications.
Related article
Yushu Technology, A-share's First Humanoid Robot Stock, Sets August 19 Listing With ~61 Billion Yuan Valuation
Scheduled for listing on the Sci-Tech Innovation Board on August 19, Yue Shu Technology will debut at 150.80 yuan per share, establishing a post-issue market capitalization of approximately 60.993 billion yuan. With a price-to-earnings ratio of 219.2
Spotify Introduces AI Persona Tag; Algorithmic Playlists Default to Exclude AI Virtual Artists
On August 11, Spotify announced the rollout of its "AI Persona" profile tagging system in mid-September. This initiative clearly identifies virtual artists generated by artificial intelligence and excludes such music from editorial and algorithmic re
OpenAI Files for IPO, Targets 2027 Listing
OpenAI’s Chief Financial Officer, Sara Ferial, announced during a Wednesday all-hands meeting that the company targets a 2027 public listing, with the possibility of an earlier debut if business momentum remains strong.Ferial emphasized that the IPO
Related Special Topic Recommendations
Comments (0)
0/500

Alibaba Cloud confirmed on July 23 that its Zhenwu M890 super node has successfully integrated with the latest flagship large model, Qwen3.8, making it available for model inference services on the Alibaba Cloud BaiLian platform. This milestone establishes the Zhenwu M890 as China’s first super node system to support the stable operation of a large model exceeding 2 trillion parameters, marking a significant breakthrough in domestic software-hardware collaboration for large model computing scheduling.
Qwen3.8, Alibaba’s newest ultra-large-scale flagship model, features up to 2.4 trillion parameters. Traditional AI clusters often face performance bottlenecks in handling high-throughput, low-latency, and high-concurrency inference demands due to limited inter-node communication bandwidth. To ensure stable and efficient operation at this scale, model parameters typically need to be split and distributed across dozens or even hundreds of high-speed connected GPU cards.
To tackle the distributed computing challenges posed by trillion-parameter models, the Alibaba Zhenwu M890 super node has optimized cluster high-speed interconnection technology and communication architecture, effectively overcoming the transmission limitations of traditional AI clusters. This ensures that large models run “stably and quickly” during inference. With this service now available on the Alibaba Cloud BaiLian platform, enterprises and developers can directly access efficient trillion-parameter inference capabilities, laying the computing foundation for the commercial deployment of next-generation ultra-large-scale AI applications.
Yushu Technology, A-share's First Humanoid Robot Stock, Sets August 19 Listing With ~61 Billion Yuan Valuation
Scheduled for listing on the Sci-Tech Innovation Board on August 19, Yue Shu Technology will debut at 150.80 yuan per share, establishing a post-issue market capitalization of approximately 60.993 billion yuan. With a price-to-earnings ratio of 219.2
Spotify Introduces AI Persona Tag; Algorithmic Playlists Default to Exclude AI Virtual Artists
On August 11, Spotify announced the rollout of its "AI Persona" profile tagging system in mid-September. This initiative clearly identifies virtual artists generated by artificial intelligence and excludes such music from editorial and algorithmic re
OpenAI Files for IPO, Targets 2027 Listing
OpenAI’s Chief Financial Officer, Sara Ferial, announced during a Wednesday all-hands meeting that the company targets a 2027 public listing, with the possibility of an earlier debut if business momentum remains strong.Ferial emphasized that the IPO











