Alibaba's Qwen team has officially launched the preview version of Qwen3.8-Max on Sunday, July 19, featuring a 2.4 trillion-parameter model that the company calls its most capable system yet. According to reports from Bloomberg, the model is described as approaching the strongest leading models, with only the model from Fable 5 US company Anthropic ahead of it. Developer Shuai Bai confirms this is the team's first multimodal model with more than 1 trillion parameters, capable of processing images, videos, and documents beyond traditional text-based AI. The company published the news through its official Qwen account on X, describing the system as 'second only to Fable 5' and placing it at the heart of the global race to develop frontier artificial intelligence technologies. The announcement was made during the country's top AI summit, the World AI Conference (WAIC) in Shanghai, underscoring how Chinese AI developers have entered the 'multi-trillion parameter' age.
Developers can now test the preview version of Qwen3.8-Max through Alibaba's coding platforms, including the Qoder platform, giving them direct access to the model's capabilities. As reported by Bloomberg, a separate Token Plan tier covers pricing for international and domestic users. The company plans to soon release the model's weights openly to a broader user base, following its open strategy that has already helped Qwen gain ground on Anthropic and OpenAI, differing sharply from closed systems such as Fable 5. According to Bloomberg, Qwen 3.8 should outperform Qwen 3.7-Max, especially in coding and complex productivity tasks such as full-stack development, data analysis, and office workflows. This approach has enabled Qwen to compete directly with the latest models developed in the US and Europe, with the preview version now made available on Alibaba's Token Plan subscription service, as well as its Qoder and QoderWork agentic platforms.
The launch comes as Alibaba seeks to disrupt Moonshot AI's momentum following the Kimi K3 model's 2.8 trillion parameter announcement. According to Bloomberg, Moonshot AI has promised open weights, but access currently runs through its chat app and API, turning interest into customers. Qwen's release could disrupt that strategy by offering more accessible, direct developer access. After reaching $300 million in annual recurring revenue in June, Moonshot plans to go public in as little as six months, making this competitive landscape particularly significant for both companies' future market positioning. These developments underscore China's rapid progress in the AI arena, with the pace of competition between Chinese companies and global players in AI recently intensifying, particularly focused on delivering powerful, highly efficient models based on massive parameter architectures and expanding access to both developers and enterprises.
Alibaba's announcement builds on a separate win with China's Cyberspace Administration approving Apple Intelligence for local release this month. According to TechNode, Alibaba will serve as a technology partner for the service alongside Baidu. The approval matters because it puts Qwen in front of millions of iPhone users in China, with Qwen models powering Apple's on-device AI features covering text, image understanding and content generation. The dual wins position Alibaba as both a model developer and infrastructure partner in China's AI market, with the company now directly competing with global AI leaders in the rapidly evolving landscape.
