AI 大模型周榜：国产模型在前端开发和视频生成领域表现亮眼

On September 7, 2026, Arena.ai released its weekly AI model rankings for week 36, covering August 31 to September 6. The rankings highlighted several new models, with multiple Chinese-developed models achieving notable results. In the frontend development category, Alibaba's Qwen3.8-max-0902 debuted at fourth place, while other models like Qwen3.8-flash-next, Tencent's Hy4-preview, and GLM-5.3-flash entered the top 15. In the AI video rankings, Alibaba's Wan3.0 made an immediate impact by securing a top-three position. The overall rankings remained stable, with Claude-fable-5 leading at 1507 points, closely followed by Claude-opus-4-6-high at 1505. New entries like Claude-fable-5.1-max and Gemini-3.8-flash-high also made their mark. In the code rankings, Anthropic models dominated the top three, while Chinese models like GLM-5.3-flash and Qwen3.8-max showed strong performance. The AI image and video rankings also saw significant participation from Chinese models, with several achieving top positions for the first time.