DeepLearning.AI· @DeepLearningAI · X·· 3 小时前AI 评分58
AI 导读
DeepLearning.AI 指出更高推理档位不等于更好结果：Claude Sonnet 5.5 在 max 档写出的代码可合并性低于低一级的 xhigh，且 token 消耗高于所有被测模型。文中引用 Terminal-Bench 4.0 数据（数据来自 Anthropic），建议读者在信任设置前先自行测试，详见 https://www.deeplearning.ai/the-batch/claude-fills-out-its-model-family
正文 · 原文
Max reasoning should mean best results, right? 🤔
Not always. Claude Sonnet 5.5 wrote less mergeable code at max than at xhigh, one level below, while using more tokens than any model measured. 🧪 Test your settings before you trust them.
📬 https://www.deeplearning.ai/the-batch/claude-fills-out-its-model-family
#DeepLearningAI #AIEngineering #AIAgents
来源：DeepLearning.AI · x.com