Alibaba Qianwen officially announced: the Qwen3.8 large model is set to be released and open-sourced, with parameter size reaching 2.4 trillion—one of the largest open-source models in China today.
Of note, the Qwen3.8-Max preview version has already been launched on Alibaba’s Token Plan, Qoder, and the QoderWork platform, allowing developers to try it out in advance.
With a 2.4T parameter scale, this generation of Qianwen is expected to see significant improvements in reasoning capability, multimodal fusion, and long-context processing. The open-sourcing strategy means developers can deploy it locally, which is an important variable for independent innovation in the AI application layer.
Competition in the AI space is shifting from the “parameter race” to the “application deployment race.” The open-sourcing of Qwen3.8 will accelerate this process.
#AI #Qwen
Of note, the Qwen3.8-Max preview version has already been launched on Alibaba’s Token Plan, Qoder, and the QoderWork platform, allowing developers to try it out in advance.
With a 2.4T parameter scale, this generation of Qianwen is expected to see significant improvements in reasoning capability, multimodal fusion, and long-context processing. The open-sourcing strategy means developers can deploy it locally, which is an important variable for independent innovation in the AI application layer.
Competition in the AI space is shifting from the “parameter race” to the “application deployment race.” The open-sourcing of Qwen3.8 will accelerate this process.
#AI #Qwen