Qwen3-Max-Thinking

Qwen has released its flagship reasoning model, Qwen3-Max-Thinking, which it says performs comparably to GPT-5.2 Thinking and Opus 4.5. Independent testing shows the model outperforms Gemini 3 Pro and GPT-5.2 on the Humanity’s Last Exam benchmark, using a test-time scaling technique that delivers efficient, high-scoring reasoning across multiple evaluations.

本文综合以下全部报道写成，而非依据任何单一来源。