.svg){kind=spacer}
Details
- OpenAI has introduced o3-pro, a more advanced version of its o3 model, now available to all ChatGPT Pro and API users as of June 10, 2025.
- Expert reviewers consistently favored o3-pro over its predecessor in areas like science, education, programming, data analysis, and writing tasks.
- The model excels under OpenAI’s rigorous “4/4 reliability” standard—requiring four consecutive correct responses—demonstrating improved consistency compared to o1-pro and o3.
- o3-pro retains robust tool integrations, including web search, Python code execution, file analysis, vision reasoning, and user-specific memory for personalized replies.
- Now set as the default for Pro and Team subscriptions, the model will expand to Enterprise and Edu next week, with no change to existing pricing.
- Safety measures follow the o3 system card guidelines, as o3-pro shares its core architecture with the previous version.
- Significant advancements are noted in advanced math, science problem-solving, and coding, as OpenAI plans to strengthen feedback mechanisms with both researchers and developers.
Impact
This move keeps OpenAI’s paid offering competitive as rivals like Anthropic’s Claude 4 and Google’s Gemini Ultra have also upgraded their features. By replacing o1-pro at no extra cost, OpenAI raises expectations for reliability and performance in the subscription market, pressuring others to match improvements while keeping prices steady. Enhanced consistency and expanded tool access further position OpenAI as a leader capable of meeting upcoming regulatory requirements and shaping the trajectory of enterprise AI adoption.
