# 弱通用AI已实现，提前一年达标

- 来源：Ethan Mollick (@emollick)
- 发布时间：2026-09-09 00:50
- AIHOT 分数：29
- AIHOT 链接：https://aihot.news/items/cmtswzjll03zxrob5bs1wghaw
- 原文链接：https://x.com/emollick/status/2097367093100343383

## AI 摘要

弱通用AI已实现（至少根据2020年设定的标准）。比近期预测提前一年 ✅勒布纳奖是弱图灵测试，GPT-4.5已等效达成 ✅Winograd被GPT-3通过 ✅SAT由GPT-4以75%通过 ✅蒙特祖玛的复仇被GPT-6攻克

## 正文

Weakly General AI achieved (at least according to the criteria set from 2020). a year before recent predictions
✅Loebner prize was a weak Turing Test, equivalent achieved by GPT-4.5
✅Winograd passed by GPT-3
✅SAT passed at 75% by GPT-4
✅Montezuma's Revenge beaten by GPT-6

### 引用推文

> fishy business：GPT-6 Astra (low) has beaten the Atari game Montezuma's Revenge, in real time, with a basic harness This presumably resolves the Metaculus question "When Will W...