OpenAI is launching a science blog where researchers can share what they have learned using artificial intelligence and the methods behind their work. The company says accelerating scientific discovery can advance its goal of ensuring that artificial general intelligence (AGI) benefits all of humanity.
The blog is intended to cover a range of disciplines, reporting new results and ongoing research from OpenAI, its collaborators and the wider scientific community. It will also offer practical examples of how the company’s models can be used for scientific computing, data analysis, literature exploration and experiment planning, including ways to check the methods and understand their limitations.
The plans also include firsthand accounts from researchers about AI’s effects on their fields, along with discussion of open questions in scientific practice and how to make the tools accessible to more people. OpenAI notes that advances, such as better medicines or new energy sources, often begin with fundamental questions whose practical significance becomes clear only years later.
Examples from mathematics and biology
The company says its models have addressed a growing range of scientific problems in recent months. In September, an internal OpenAI system produced a proof addressing the Navier–Stokes existence and smoothness problem, a longstanding question involving the equations that describe fluid motion. In early October, the company published a range of mathematical results, including one on the quasi-Riemann hypothesis that sharpens estimates of how many prime numbers lie below a given number.
Later in October, OpenAI reported results in mathematics, computer science and physics. They concerned the concentration of waves, the limits of efficient computation and the mathematical foundations of interlacing quantum field theory.
In biology, researchers are using AI to analyze complex datasets, connect findings across fields and develop hypotheses that can be tested in the laboratory. OpenAI described work with PerturbAI and Scripps Research in which thousands of agents powered by GPT-Rosalind analyzed existing experimental data from millions of mouse brain cells in 32 minutes. By connecting the effects of gene disruptions with scientific literature, they identified possible biological mechanisms for researchers to test.
Other researchers are developing their own methods. César de la Fuente’s team uses ChatGPT and Codex to build algorithms that identify potential antibiotic molecules. The team then makes and tests those molecules in the laboratory.
Changes to scientific practice
The new capabilities also affect how research is conducted. In mathematics, systems that generate proofs raise questions about how new arguments can be checked, explained and incorporated into the field. In experimental sciences, exploring more hypotheses creates opportunities but can also make it harder to decide which experiments to pursue.
OpenAI says researchers across disciplines must also consider implications for training, attribution and the skills their work requires. The effects will vary by field, and scientific communities should have a role in shaping how the tools are used.



