OpenAI o3

Last updated

o3
Developer(s) OpenAI
Initial release
  • o3-mini: January 31, 2025 (2025-01-31)
  • o3: April 16, 2025 (2025-04-16)
o3-pro: June 10, 2025 (2025-06-10)
Predecessor OpenAI o1
Successor
Type

OpenAI o3 is a reflective generative pre-trained transformer (GPT) model developed by OpenAI as a successor to OpenAI o1 for ChatGPT. It is designed to devote additional deliberation time when addressing questions that require step-by-step logical reasoning. [1] [2] On January 31, 2025, OpenAI released a smaller model, o3-mini, [3] followed on April 16 by o3 and o4-mini. [4]

Contents

History

The OpenAI o3 model was announced on December 20, 2024. It was called "o3" rather than "o2" to avoid trademark conflict with the mobile carrier brand named O2. [1] OpenAI invited safety and security researchers to apply for early access of these models until January 10, 2025. [5] Similarly to o1, there are two different models: o3 and o3-mini. [3]

On January 31, 2025, OpenAI released o3-mini to all ChatGPT users (including free-tier) and some API users. OpenAI describes o3-mini as a "specialized alternative" to o1 for "technical domains requiring precision and speed". [6] o3-mini features three reasoning effort levels: low, medium and high. The free version uses medium. The variant using more compute is called o3-mini-high, and is available to paid subscribers. [3] [7] Subscribers to ChatGPT's Pro tier have unlimited access to both o3-mini and o3-mini-high. [6]

On February 2, OpenAI launched OpenAI Deep Research, a ChatGPT service using a version of o3 that makes comprehensive reports within 5 to 30 minutes, based on web searches. [8]

On February 6, in response to pressure from rivals like DeepSeek, OpenAI announced an update aimed at enhancing the transparency of the thought process in its o3-mini model. [9]

On February 12, OpenAI further increased rate limits for o3-mini-high to 50 requests per day (from 50 requests per week) for ChatGPT Plus subscribers, and implemented file/image upload support. [10]

On April 16, 2025, OpenAI released o3 and o4-mini, a successor of o3-mini. [4]

On June 10, OpenAI released o3-pro, which the company claims is its most capable model yet. [11] OpenAI stated: "We recommend using it for challenging questions where reliability matters more than speed, and waiting a few minutes is worth the tradeoff". [12]

Capabilities

Reinforcement learning was used to teach o3 to "think" before generating answers, using what OpenAI refers to as a "private chain of thought". [13] This approach enables the model to plan ahead and reason through tasks, performing a series of intermediate reasoning steps to assist in solving the problem, at the cost of additional computing power and increased latency of responses. [14]

o3 demonstrates significantly better performance than o1 on complex tasks, including coding, mathematics, and science. [1] OpenAI reported that o3 achieved a score of 87.7% on the GPQA Diamond benchmark, which contains expert-level science questions not publicly available online. [15]

On SWE-bench Verified, a software engineering benchmark assessing the ability to solve real GitHub issues, o3 scored 71.7%, compared to 48.9% for o1. On Codeforces, o3 reached an Elo score of 2727, whereas o1 scored 1891. [15]

On the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI) benchmark, which evaluates an AI's ability to handle new logical and skill acquisition problems, o3 attained three times the accuracy of o1. [1] [16]

See also

References

  1. 1 2 3 4 Knight, Will (December 20, 2024). "OpenAI Upgrades Its Smartest AI Model With Improved Reasoning Skills". Wired .
  2. Metz, Cade (December 20, 2024). "OpenAI Unveils New A.l. That Can 'Reason' Through Math and Science Problems". The New York Times .
  3. 1 2 3 Franzen, Carl (January 31, 2025). "It's here: OpenAI's o3-mini advanced reasoning model arrives to counter DeepSeek's rise". VentureBeat. Retrieved February 1, 2025.
  4. 1 2 "Introducing OpenAI o3 and o4-mini". openai.com. Retrieved April 16, 2025.
  5. "Early access for safety testing". OpenAI. December 20, 2024.
  6. 1 2 "Introducing OpenAI O3 Mini". OpenAI. February 13, 2025. Retrieved February 13, 2025.
  7. "OpenAI o3-mini". OpenAI. January 31, 2025. Retrieved February 3, 2025.
  8. Ha, Anthony (February 3, 2025). "OpenAI unveils a new ChatGPT agent for 'deep research'". TechCrunch. Retrieved February 4, 2025.
  9. Wiggers, Kyle (February 6, 2025). "OpenAI now reveals more of its o3-mini model's thought process". TechCrunch. Retrieved February 7, 2025.
  10. @OpenAI (February 13, 2025). "Two updates you'll like— OpenAI o1 and o3-mini now support both file & image uploads in ChatGPT. We raised o3-mini-high limits by 7x for Plus users to up to 50 per day" (Tweet). Retrieved February 13, 2025 via Twitter.
  11. Wiggers, Kyle (June 10, 2025). "OpenAI releases o3-pro, a souped-up version of its o3 AI reasoning model". TechCrunch. Retrieved June 11, 2025.
  12. Washenko, Anna (June 6, 2025). "OpenAI adds the o3-pro model to ChatGPT today". Engadget. Retrieved June 11, 2025.
  13. OpenAI O3 Mini System Card (PDF) (Report). OpenAI. January 13, 2025. Retrieved February 13, 2025.
  14. Zeff, Maxwell; Wiggers, Kyle (December 20, 2024). "OpenAI announces new o3 models". TechCrunch . Retrieved December 22, 2024.
  15. 1 2 Franzen, Carl; David, Emilia (December 20, 2024). "OpenAI confirms new frontier models o3 and o3-mini". VentureBeat . Retrieved December 26, 2024.
  16. Hsu, Jeremy (December 20, 2024). "OpenAI's o3 model aced a test of AI reasoning – but it's still not AGI". New Scientist . Retrieved December 22, 2024.