Particle.news
Download on the App Store

OpenAI Declares ‘Code Red’ To Accelerate ChatGPT as Ads, Agents Are Put on Hold

OpenAI shifts resources to ChatGPT quality as Google’s Gemini 3 posts strong benchmarks.

Overview

  • An internal memo reported by The Information and the Wall Street Journal says Sam Altman ordered a company-wide push to improve ChatGPT’s speed, reliability, personalization and breadth of answers, with daily coordination calls and temporary team transfers.
  • OpenAI plans to release a new reasoning model next week, which the memo claims outperforms Google’s Gemini 3 in internal tests, a performance assertion that has not been publicly validated.
  • Work on advertising in ChatGPT is paused despite recent beta code references in the Android app, and projects including shopping and health agents and the personal assistant Pulse are delayed.
  • Google’s Gemini 3, launched in mid-November, has led recent benchmarks such as Humanity’s Last Exam and reported 650 million monthly users, drawing endorsements from high-profile users like Marc Benioff.
  • OpenAI says ChatGPT remains the leading assistant with about 800 million weekly users, while Altman warned staff of possible temporary economic headwinds from renewed competition.