Alibaba has released the upgraded version of Qwen3, with performance surpassing leading proprietary models like Kimi-K2.

According to the Qwen Q&A: Today, we have updated the flagship version of the Qwen3 model, releasing an updated version of the Qwen3-235B-A22B-FP8 non-thinking mode, named Qwen3-235B-A22B-Instruct-2507-FP8. The new Qwen3 model has significantly improved general abilities, including instruction following, logical reasoning, text understanding, mathematics, science, programming, and tool usage. It has performed exceptionally well in various evaluations like GQPA, AIME25, LiveCodeBench, Arena-Hard, BFCL, surpassing top open-source models like Kimi-K2, DeepSeek-V3, as well as leading closed-source models like Claude-Opus4-Non-thinking.