精选理由
DeepSeek新模型在Text Arena拿1465分，和GLM-5.1、GPT-5.6 Terra一个水平，开源阵营排第五。
DeepSeek-V4-Pro (Max)在Text Arena开源模型榜单中位列第五，AutoEval得分为1465。该分数与GLM-5.1的1467分、GPT-5.6 Terra (xHigh)的1464分以及Grok 4.6 (High)的1464分基本持平。该排名来自@deepseek_ai的官方发布，针对的是文本竞技场中的开源模型。
原文 · lmarena.ai
In the Text Arena: DeepSeek-V4-Pro (Max) by @deepseek_ai is coming in around #5 among open models in...
In the Text Arena: DeepSeek-V4-Pro (Max) by @deepseek_ai is coming in around #5 among open models in the Text Arena! With 1465pts (AutoEval), it’s on par with GLM-5.1 (1467 pts), GPT-5.6 Terra (xHigh) (1464 pts) and Grok 4.6 (High) (1464 pts). 💬 1 🔄 0 ❤️ 9 👀 805 📊 2 ⚡