OpenAI has brought back GPT-4o as the default model for its ChatGPT paying users. If you’re on a Plus, Pro, Team, Enterprise, or Edu plan, you’ll now find GPT-4o readily available – a move that comes after some early hitches with GPT-5. The initial launch of GPT-5 on August 7 introduced a variety of modes and variants, which, despite their promise, led to some misrouted prompts and inconsistent performance.
The updated interface now includes a convenient ‘Show additional models’ setting. This means beyond GPT-4o, you can effortlessly access GPT-4.1, o3, and o4-mini, each designed with enhanced reasoning capabilities in mind. Sam Altman, OpenAI’s CEO, shared on X that if GPT-4o ever goes away again, users will get plenty of advance notice.
GPT-5 still holds its own appeal with fresh features like ‘Auto’, ‘Fast’, and ‘Thinking’ modes. The ‘Thinking’ option supports a massive 196,000-token context window but comes with a weekly limit of 3,000 messages before switching to a leaner version – GPT-5 Thinking mini. Meanwhile, due to high GPU expenses, GPT-4.5 remains exclusive to Pro users.
Looking ahead, OpenAI is tweaking GPT-5 to make it friendlier and less divisive compared to GPT-4o. They’re even considering user customisation to address the emotional bonds many of us have formed with previous models. If you’ve ever felt the frustration of abrupt changes, it’s reassuring to know that these updates are designed to smooth your experience and maintain a reliable, powerful tool for daily tasks.
