OpenAI has released a new product called OpenAI o3-pro on June 10, 2025.
It is available in ChatGPT Pro / Team model picker and API endpoint o3-pro-2025-06-10.
The pricing for OpenAI o3-pro is $20 per million input tokens and $80 per million output tokens, which is higher than the standard o3 pricing of $2 and $8 respectively.
Key upgrades of OpenAI o3-pro include a longer 'reflection' loop, built-in Web search, file & image analysis, Python execution, and memory personalization.
There are initial limits to OpenAI o3-pro such as no image generation or Canvas, a temporary-chat bug, and an API launch cap of around 20k TPM per user.
A positive sentiment about OpenAI o3-pro is that it can handle various tasks efficiently; for example, producing a full research memo in ten minutes, a task that previously took half a day.
Reviewers consistently prefer o3-pro in areas like science, programming, business, and writing.
Some neutral thoughts about OpenAI o3-pro include acknowledging better quality but facing higher latency and cost compared to the standard version.
It is suggested to route only critical calls to o3-pro to manage costs effectively.
Users are advised to choose the model based on the use-case, as models like Claude or Gemini may be more cost-effective for bulk tasks.