Arena Blind Test Rankings: Chinese AI Models Hold Steady, Kimi Enters Coding Top 10

The Arena platform has released its Week 42 AI model blind test rankings for 2026. The list is based on anonymous user votes and ELO scores, showing what people prefer. The top spots stayed stable this week.

In the overall list, Google's gemini-4-argon-high holds first place with 1525 points, and Anthropic's claude-opus-5.5-high comes second. Chinese models hold steady: Moonshot's kimi-k3-max ranks 16th, while Alibaba's qwen3.8-max, Xiaomi's mimo-v2.6-pro, and Zhipu's glm-5.3-max land 22nd to 27th.

On the coding list, kimi-k3-max ranks 10th, reaching the top ten for the first time, with Xiaomi's mimo-v2.6-pro right behind at 11th. On the frontend list, mimo-v2.6-pro debuts at 19th, and several Chinese models stay in the upper-middle range.