OpenAI announced a set of commitments to support American scientific research, primarily through a partnership with the U.S. Department of Energy‘s Genesis Mission. The initiative aims to integrate frontier AI into national scientific infrastructure, helping to compress decades of progress into years. OpenAI will provide $4 million in Codex access to approximately 2,000 Genesis researchers at National Laboratories and universities, and commit $3 million in API support for two large-scale scientific campaigns. Participating researchers can access up to $10 million in API usage for $2.5 million spent. Additionally, selected national-laboratory researchers on eligible biology projects will get access to GPT‑Rosalind’s specialized bioscience capabilities. Trusted leaders will receive early access to selected models, and cybersecurity researchers will get expanded access to advanced cyber capabilities for defensive research. These commitments build on prior work, including an AI Jam Session with over 1,000 scientists from nine National Laboratories, and deployment of advanced reasoning models on the Venado supercomputer at Los Alamos. The initial large-scale campaigns will pursue breakthroughs for high-temperature superconductors and create an ‘Atlas of the Machine-Accessible Frontier’ to identify problems where AI can already drive scientific advances. OpenAI frames this as a long-term collaborative effort requiring coordination among government, labs, universities, and industry to build sustained national capacity for applying frontier intelligence to important problems.

The announcement explicitly notes that no single entity can realize this opportunity alone, positioning the company’s role as contributing frontier capabilities within a broader ecosystem where researchers define questions, choose methods, challenge outputs, and validate results. The Genesis Mission itself aims to double the productivity and impact of American research within a decade by integrating AI with federal data, computing, experimental facilities, and expert teams.

Notable boundaries and limits stated in the source include that the collaboration is focused on national-security and public-sector scientific research, access grants specify caps ($4M Codex, $3M API campaigns, $10M total API for $2.5M spent), and bioscience work with Los Alamos explicitly includes studying risks as systems become capable of ‘consequential scientific work.’ The programs are structured around trust and security, with selected researchers in designated areas receiving specialized access rather than universal open access to OpenAI models for all national laboratory scientists. The source provides no timeline, deployment details, or performance metrics from the announced efforts, and does not claim that OpenAI models provide the only path or are superior to alternatives for these scientific goals. OpenAI characterizes its contribution as ‘frontier capabilities’ within a larger ecosystem of ‘people, supercomputers, simulations, and facilities’ that already make American science exceptional, and the ‘long-term commitment’ language implies that measurable outcomes from these campaigns are prospective rather than already achieved.