OpenAI · OpenAI (GPT / ChatGPT)
o3
Activeo-series reasoning model for hardest math/science; retiring from ChatGPT Aug 26, 2026 (API still active).
About this version
The o3 model is the third iteration of OpenAI’s chatgpt family, positioned in the reasoning tier and released on 1 April 2025 with a knowledge cutoff of January 2025. It features an expanded context window of 200 000 tokens and is targeted at solving the most challenging math and science problems; it will be retired from the ChatGPT interface on 26 August 2026 while remaining available via API.
- API string
- o3
- Context window
- 200K ctx
- Knowledge cutoff
- Jan 1, 2025
- Tier
- Reasoning
- Released
- Apr 1, 2025
Release & change timeline
2 events- Deprecation
o3 retiring from ChatGPT
o3 retires from ChatGPT August 26, 2026; API access continues.
- Release
o3 released
New reasoning-model generation; topped AIME 2024 and 2025.
Report an issue with OpenAI (GPT / ChatGPT)
Posts to your status feed
Pick the closest match below, edit the body, and post. Your report carries the #chatgpt tag automatically so it surfaces here + in the trending-tags rail. Reporting also follows OpenAI (GPT / ChatGPT) so you’ll get status updates.
Part of the OpenAI (GPT / ChatGPT) version timeline. Catch Me Up — everything that changed across AI & smart home ↗