All corrections
www.oneusefulthing.org/p/sign-of-the-future-gpt-55
1 correction found
1
Claim
OpenAI’s first reasoning model, o3
Correction
o3 was not OpenAI’s first reasoning model. OpenAI introduced o1-preview as the first model in its reasoning series in September 2024, months before o3 launched in April 2025.
Full reasoning
OpenAI’s own documentation says the o1 line—not o3—was the first reasoning series.
- In OpenAI’s September 12, 2024 announcement for o1-preview, the company says: “Today, we are releasing the first of this series in ChatGPT and our API.”
- OpenAI’s API docs also describe o1-preview as “our first o-series reasoning model.”
- By contrast, OpenAI’s April 16, 2025 o3 launch post calls o3 and o4-mini “the latest in our o-series”, which places o3 after o1 chronologically.
So the post’s wording is historically incorrect: o3 was a later reasoning model, not OpenAI’s first one.
3 sources
- Introducing OpenAI o1-preview | OpenAI
Today, we are releasing the first of this series in ChatGPT and our API.
- o1 Preview Model | OpenAI API
Preview of our first o-series reasoning model.
- Introducing OpenAI o3 and o4-mini | OpenAI
Today, we’re releasing OpenAI o3 and o4-mini, the latest in our o-series of models trained to think for longer before responding.