Thousand Questions 3.5 flagship version preview model appeared, surpassing GPT5.4 and Claude4.6 on the list.

Qwen3.5 flagship preview model debuts on the list surpassing GPT5.4, Claude4.6. The globally renowned large model blind test list LMArena has updated its rankings, with Alibaba's latest flagship model preview version Qwen3.5-Max-Preview making its first appearance, scoring 1464 points, surpassing overseas models such as GPT5.4, Grok4.1, as well as domestic models such as DouBao2.0, GLM5, Kimi2.5. There are reports that the official version of Qwen3.5-Max may be released soon, with its capabilities continuing to evolve and upgrade. In addition, LMArena ranks global large model institutions based on their strongest models, with 5 Chinese companies breaking into the top ten, with Alibaba ranking in the top five globally and number one in China, while companies such as ByteDance, Zhupu, Dark Side of the Moon, and Baidu also entering the global top ten.