OpenAI first previewed the o3 and o4-mini models during their 'shipmas' event in December 2024, with early testers gaining access in January 2025.

Additionally, these models are expected to generate original research ideas in complex fields like drug discovery and materials science, which have traditionally been dominated by PhD-level experts.

These models are part of OpenAI's ongoing efforts to enhance artificial intelligence applications in practical, real-world scenarios.

OpenAI is expected to charge $20,000 per month for access to these advanced AI models, a significant increase compared to the $200 monthly fee for ChatGPT Pro.

OpenAI is preparing to launch its advanced AI models, o3 and o4-mini, which aim to revolutionize scientific innovation and research.

Part of the GPT-4.1 family, these reasoning models are specifically trained to provide higher quality responses for STEM tasks.

Utilizing large datasets and advanced machine learning techniques, the models aim to generate novel experimental ideas and solutions.

They could be particularly effective in fields such as nuclear fusion research and pathogen detection, pushing the boundaries of current scientific methodologies.

The o3 and o4-mini models are designed to analyze complex questions, produce testable hypotheses, identify gaps in research, and design experiments, moving beyond basic data summarization.

The ability of these models to create viable experiments marks a significant advancement, as such tasks have traditionally been performed by humans.

While AI will not replace scientists immediately, its evolving capabilities could transform the future of scientific discovery by thinking and operating like researchers.

Although the release date is currently undetermined, it could happen as soon as the week of April 15, 2025.