OpenAI o4-mini

Last updated
o4-mini
Developer(s) OpenAI
Initial releaseApril 16, 2025 (2025-04-16)
Predecessor OpenAI o3-mini
Successor GPT-5 mini
Type

OpenAI o4-mini is a generative pre-trained transformer model created by OpenAI. On April 16, 2025, the o4-mini model was released to all ChatGPT users (including free-tier users) as well as via the Chat Completions API and Responses API. [1] [2] [3] [4]

Contents

Additionally, OpenAI introduced the o4-mini-high model, which was made available exclusively to paid-tier ChatGPT users. The high model offers more advanced features, including higher response accuracy and faster processing times. [5]

Unlike earlier models, o4-mini is capable of processing both text and images. It also allows to perform tasks like analyzing whiteboard sketches during its "chain-of-thought" phase. [6]

o4-mini API providers says that it's designed to enhance decision-making across sectors by enabling utilities to forecast demand and analyze infrastructure data, supporting healthcare through extraction and interpretation of medical records and diagnostics, and assisting financial institutions with real-time regulatory compliance and risk assessment through automated document analysis and data processing. [7] A study shows that the o4-mini model has significantly improved its ability to complete tasks in the field of chemistry compared to the GPT-3 model, but there is still a gap compared to other models (Such as Deepseek-R1 model). [8]

See also

References

  1. "Introducing OpenAI o3 and o4-mini". OpenAI. Retrieved 17 April 2025.
  2. Mauran, Cecily (16 April 2025). "OpenAI announces o3 and o4-mini reasoning models for ChatGPT (updated)". Mashable. Retrieved 17 April 2025.
  3. Nuñez, Michael (16 April 2025). "OpenAI launches o3 and o4-mini, AI models that 'think with images' and use tools autonomously". VentureBeat. Retrieved 17 April 2025.
  4. Metz, Cade (2025-04-16). "OpenAI Unveils New 'Reasoning' Models o3 and o4-mini". The New York Times. ISSN   0362-4331 . Retrieved 2025-04-17.
  5. "OpenAI bringt o3 und o4-mini raus". heise online (in German). 2025-04-17. Retrieved 2025-04-17.
  6. Zeff, Maxwell (2025-04-16). "OpenAI launches a pair of AI reasoning models, o3 and o4-mini". TechCrunch. Retrieved 2025-04-17.
  7. AI/ML, API (2025-04-17). "Build With OpenAI o4-mini". AIML API. Retrieved 2025-04-17.
  8. Cui, Dong-xu; Long, Shi-yu; Li, Qiao (2025-06-02), Can reasoning power significantly improve the knowledge of large language models for chemistry? --Based on conversations with Deepseek and ChatGPT, American Chemical Society (ACS), doi:10.26434/chemrxiv-2025-85wpr-v3 , retrieved 2025-08-12