On 12 September 2024 OpenAI released o1-preview and o1-mini. It said it had trained them to think longer before answering, and that on a qualifying exam for the International Mathematics Olympiad GPT-4o solved 13% of problems while the reasoning model scored 83%. It reset the counter to 1 and named the series OpenAI o1.
For someone nervous about “versions”: this was not simply the next GPT number. OpenAI said it was starting a new line. Results on harder maths and code improved. Everyday chat often still suited GPT-4o.
Intelligence Age Gulf wrote this note on 30 September 2026.
Source. OpenAI, Introducing OpenAI o1-preview, 12 September 2024.