Bỏ qua đến nội dung chính
Back to home
AI tools-ai Tech 2 min read

Alibaba Launches Qwen3.8-Max to Challenge OpenAI With 2.4 Trillion Parameters

Chinese tech giant Alibaba has unveiled Qwen3.8-Max, a massive open-weight AI model boasting 2.4 trillion parameters, designed to compete directly with leading U.S. rivals.

Tier 1 · sources 99% confidence Reviewed
📚 Aggregated from 2 sources The Verge The Decoder

Chinese tech giant Alibaba has officially unveiled Qwen3.8-Max, its largest and most powerful artificial intelligence model to date. According to reports released on August 3, 2026, this model is designed to handle complex, multi-day tasks and is expected to compete directly with the most advanced AI systems from OpenAI and Anthropic.

Detailed Developments

The launch of Qwen3.8-Max marks a strategic move by Alibaba in the race for global AI leadership. According to a report by 'The Verge', the Chinese tech giant claims the new model delivers exceptional performance, not only competing with top U.S. labs but also outperforming domestic rivals like Moonshot AI's Kimi K3. While Alibaba announced it has begun making the model widely available, 'The Decoder' reported that the development team plans to release the model weights next week. This promise is poised to give a significant boost to the global open-source AI research community.

Technical & Technology Analysis

Technically, Qwen3.8-Max features a massive architecture with a scale of 2.4 trillion parameters. A key highlight of this open-weight model is its ability to handle long-horizon tasks autonomously and continuously over several days. According to technical details shared by 'The Decoder', the system can independently perform highly complex tasks, such as replicating scientific research papers or even designing semiconductor chips without continuous human intervention. Sustaining autonomous operations over long periods requires breakthrough optimizations in both model architecture and hardware memory management.

Expert Opinions & Insights

Industry experts view Alibaba's decision to release open weights for a 2.4-trillion-parameter model as a bold strategy. Many analysts believe this move aims to disrupt the monopoly of proprietary U.S. models like GPT-4 or Claude 3. However, observers remain cautious regarding Alibaba's performance claims. The industry must wait for independent community benchmarks following the weight release next week to verify if Qwen3.8-Max truly delivers on its promises.

Impact & Future Outlook

The arrival of Qwen3.8-Max indicates that the technological gap between U.S. and Chinese AI studios is narrowing rapidly. For the tech community and developers in Vietnam, accessing an open-weight model of this scale opens up massive opportunities for customization and real-world application without relying on expensive, closed-source APIs. The near-future AI race will certainly no longer be the exclusive playground of Silicon Valley giants, as Asian players accelerate rapidly with ambitious open-source strategies.