今天分别用 Codex (GPT-6-Astra) 和 Grok Bot 研究一个问题
Codex 用的更早，所以它的上下文其实更完整。但今天用下来，Grok Bot 的回答明显更有参加价值，Codex 更空泛套话，没有深入到具体的问题背景里。
不确定 Grok Bot 背后默认模型现在是什么，但确实好用。我还没用 Claude Code 做对比，因为。。以前的账号都封了 😂
Shao compares Codex (GPT-6-Astra) and Grok Bot on one research question
AISummary
Meng Shao says he used Codex with GPT-6-Astra and Grok Bot on the same research question, and found Grok Bot's answer more useful. He says Codex gave generic answers without the specific background, though it had more context since he used it earlier. He does not know which default model Grok Bot currently uses and has not compared it with Claude Code.
Post on XView on X
meng shaoVerified on X
@shao__meng
Source: meng shao · x.comPublished