Qwen takes 3rd place globally.

The internationally renowned large-scale model evaluation Chatbot Arena has announced its rankings, with Qwen3-235B-A22B-Instruct-2507 scoring 1433 points, surpassing top closed-source models Grok4, Claude4, and GPT4.1, ranking third globally on the overall leaderboard. Chatbot Arena uses a blind evaluation mechanism and is one of the most influential lists in the AI large model field. The 1433 points achieved by Qwen3 in this competition are the highest in the history of global open-source and Chinese large models. Additionally, Qwen3 also secured the top spot globally in 5 key ability sub-items, including mathematics, coding, complex prompts, long-text retrieval, and command compliance.