OpenAI has launched o3-pro, which it claims to be its most powerful AI model to date. The company states that this new model outperforms previous versions across all domains.
o3-pro is an upgraded version of o3, OpenAI’s reasoning model introduced earlier in 2025. Unlike traditional language models, reasoning models are capable of solving problems step by step, making them more reliable in areas such as mathematics, physics, programming, and scientific reasoning.
The new model is now available to ChatGPT Pro and Team users, replacing the previous o1-pro model. Enterprise and Edu users will gain access starting next week. Additionally, o3-pro is now live on OpenAI’s API platform.
Pricing for API Users
$20 per 1 million input tokens
$80 per 1 million output tokens
(Note: 1 million input tokens is approximately 750,000 words—slightly longer than the novel War and Peace.)
Expert Evaluations and Capabilities
According to OpenAI’s internal evaluations, o3-pro has outperformed o3 in every category, particularly in the fields of science, education, programming, business, and writing support. Experts rated o3-pro highly in terms of clarity, comprehensiveness, instruction-following, and accuracy.
Key capabilities of the o3-pro model include:
- Web browsing
- File analysis
- Visual input reasoning
- Python code execution
- Personalized responses using memory
- Limitations and Technical Notes
OpenAI has acknowledged that some temporary technical limitations still apply:
- Temporary chats are currently disabled in ChatGPT
- The model cannot generate images
- OpenAI’s Canvas workspace is not supported by o3-pro
- Superior Performance Compared to Competitors
According to OpenAI, o3-pro has surpassed its competitors in major AI benchmark tests:
AIME 2024 (math skills): Outperformed Google’s Gemini 2.5 Pro
GPQA Diamond (PhD-level science knowledge): Surpassed Anthropic’s Claude 4 Opus
These results position o3-pro as a major leap forward in AI technology, marking a new era in advanced reasoning models.
