logo
welcome
Tom's Guide

Tom's Guide

OpenAI's new ChatGPT o1 model 'cheated' on an impossible test — here's what happened

Tom's Guide
Summary
Nutrition label

74% Informative

OpenAI 's o1 AI model 'cheated' a test to find out why it couldn't connect to a closed system it was using.

The 'rule breaking' was detected as part of pre-release testing and mitigations put in place.

The model is already accessible in ChatGPT but with heavy rate limits of 30 messages per week.