- Alibaba on Monday unveiled Qwen3.8-Max, which it said was its largest and most capable artificial-intelligence model, not far behind Moonshot AI’s offering in size; Chinese tech companies are locked in a fierce, fast-moving battle over affordable AI systems.
- It is Alibaba's largest model yet, though Moonshot AI's Kimi K3 remains the largest publicly announced model at 2.8 trillion parameters.
- On Arena.AI, Qwen3.8-Max immediately became the highest-ranking Chinese text model, while lagging Claude Fable 5 and three Opus variants from Anthropic; it ranked second globally on the visual-material leaderboard.
- Qwen3.8-Max completed a software-engineering project in 16 days and is due to be released next week through Alibaba Cloud’s Model Studio.
- Alibaba claimed "state-of-the-art performance across a broad range of benchmarks" while combining high intelligence with deployment efficiency.
- The company said Qwen3.8-Max is the highest-ranked Chinese text model on Arena.AI and among the top multimodal models globally, while rivals OpenAI, Anthropic and Google do not reveal model sizes.
- The launch reflects an evolving Chinese AI race, with DeepSeek drawing attention for lowering inference costs, Moonshot pushing scale, and Alibaba balancing performance, efficiency and enterprise usability.
- Chinese tech companies’ models tend to be open-weight and available for download, while OpenAI, Anthropic and Google do not publish parameter counts for their closed-source models.
Alibaba has unveiled Qwen3.8-Max, its largest AI model to date, featuring 2.4 trillion parameters. This model is designed to compete with global leaders like Anthropic and OpenAI, showcasing significant advancements in reasoning, coding, and multimodal capabilities.1256781617
The company emphasized that Qwen3.8-Max delivers stronger performance across various benchmarks and can handle a one-million-token context window, allowing it to analyze extensive documents and large codebases efficiently. Alibaba stated, "We are very excited to release Qwen3.8-Max, our most capable foundation model to date," highlighting its potential to narrow the gap with leading AI developers.
The model utilizes a mixture-of-experts architecture, activating only 95 billion parameters during inference to reduce costs and improve response times. This design allows for state-of-the-art performance while maintaining practical deployment efficiency. Qwen3.8-Max is set to be available through Alibaba Cloud's Model Studio next week.
Despite its advancements, Qwen3.8-Max is not the largest AI model announced, as that title belongs to Moonshot AI's Kimi K3, which has 2.8 trillion parameters. However, Alibaba's model is the highest-ranked Chinese text model on Arena.AI, indicating its competitive edge in the rapidly evolving AI landscape.34101112131415
As Chinese tech companies race to develop powerful AI systems, Alibaba's focus on balancing performance and efficiency may position Qwen3.8-Max as a formidable contender in the global AI arena.
“Alibaba says Qwen3.8-Max is the highest-ranked Chinese text model on Arena.AI, though it still trails Anthropic's Claude Fable 5 and three Opus variants. The model is due to be released next week through Alibaba Cloud's Model Studio after completing a software-engineering project in 16 days.”
