OpenAI pledges $5 million to Simons Foundation mathematics initiative

OpenAI says its $5 million contribution is going to a new Simons Foundation initiative focused on helping mathematicians examine, understand and build on results produced by AI.[1] The foundation will manage and award OpenAI’s contribution.[1] OpenAI says the funding will support workshops, conferences and special programs.[1] The initiative’s stated aim aligns with OpenAI’s broader goal of turning AI progress into deeper human understanding and new discoveries.[1]

Alongside the contribution, OpenAI says it will provide its models and computing resources to researchers supported by the Simons Foundation.[1] The foundation says its own commitment is $15 million for workshops and initiatives, and that foundation-supported researchers will receive free AI computing resources.[1] These are separate statements about the two organizations’ roles: OpenAI is contributing $5 million and providing resources, while the foundation manages and awards that contribution.[1]

OpenAI says the initiative will also support work on results it recently published.[1] The company says those mathematics results were produced by an internal frontier model and published in a GitHub repository with protocols for paper revisions and citations.[2] OpenAI says the repository includes 10 summaries of the model’s reasoning, compute estimates and statistics on attempted problems.[2] It also says the average result used computing equivalent to roughly three hours of ChatGPT Pro thinking.[2] Separately, OpenAI released more than 700 mathematical preprints entirely generated by AI on October 6, according to Nature.[3]

OpenAI posted its announcement on X on October 15, 2026.[1] Responses cited in the evidence included a commenter questioning why mathematicians were receiving special treatment and another criticizing the decision to give money to mathematicians.[1] OpenAI’s announcement frames the initiative as support for mathematicians to scrutinize and build on AI-produced work, rather than as a claim that AI results alone constitute deeper human understanding.[1]