According to online reports, Alibaba Tongyi Qianwen released a new version of the Qwen3 series models early this morning, including 2 MoE models and 6 dense models, with parameters ranging from 0.6B to 235B. Compared with other top models such as DeepSeek-R1, o1, o3-mini, Grok-3 and Gemini-2.5-Pro, the flagship model Qwen3- 235B-A22B has achieved competitive results in benchmarking such as coding, mathematics, and universal capabilities. Tongyi Qianwen also said that the reasoning capabilities of the Qwen3 model have been greatly improved, reaching the SOTA level of the same scale industry in evaluations such as mathematics, code and logical reasoning.