Artificial Analysis Ranking: Alibaba's Qwen3.8-Max Tops the World in Agentic Capabilities
Alibaba/Qwen
Anthropic
OpenAI
Moonshot AI
On August 6, Artificial Analysis released its latest rankings, with Alibaba's Qwen3.8-Max taking first place globally in the Agentic Index, surpassing Claude Opus5 and GPT5.6. The model scored 55.4 points, outperforming the previous best for Chinese models, Kimi K3's 50.1 points. This achievement marks a significant milestone as the top spot was previously dominated by Western models.
Artificial Analysis announced its latest ranking list on August 6, in which Alibaba's Qwen3.8-Max topped the Agentic Index, surpassing Claude Opus5 and GPT5.6 to rank first globally. Agentic capabilities represent the potential of AI to call tools for solving more complex problems, which is key to AI implementation in professional scenarios. This field had long been monopolized by models like Claude and GPT, and the previous best performance for a Chinese model was Kimi K3 with a score of 50.1. Alibaba's Qwen3.8 achieved first place with a score of 55.4 in the Agentic capability ranking by Artificial Analysis. This article is provided by Alibaba and republished with permission by QbitAI, and the views belong to the original authors.
Source: QbitAI 量子位 —
original