OpenAI’s o3-pro: New AI Model Outperforms Gemini and Claude in Key Academic Tests

OpenAI has launched o3-pro, a new AI model designed for enhanced reasoning capabilities. Available from June 10, o3-pro is currently accessible to ChatGPT Pro and Team users, succeeding the o1-pro model, with enterprise and education users gaining access next week. This model builds upon the original o3 model, offering greater reliability in areas such as coding, mathematics, and science. OpenAI highlights its effectiveness in long-form tasks where precision is crucial. Although slower, o3-pro delivers better outcomes. Internal assessments showed a preference for o3-pro across scientific, programming, business, and writing tasks. In benchmark tests, o3-pro excelled, surpassing Google’s Gemini 2.5 Pro on the AIME 2024 math benchmark and Anthropic’s Claude 4 Opus on the GPQA Diamond test, which measures PhD-level science comprehension. API pricing is set at $20 per million input tokens and $80 per million output tokens. Technical limitations currently prevent temporary chats, and image generation is not supported. ChatGPT Pro users can now directly select o3-pro, making it ideal for in-depth research, coding assistance, and content creation where accuracy is essential. o3-pro incorporates the same safety features as o3.