Qwen has become one of the most talked-about AI model families of 2026, and the claim that it “beats Gemini and ChatGPT in coding” is circulating widely. The claim traces back to a real result, just not a coding one. Qwen3-Max-Thinking outperformed Gemini 3 Pro and GPT-5.2 on Humanity’s Last Exam, a demanding general reasoning and knowledge benchmark, when paired with search tools.
Model