An internal version of our next major model produced 10 new results on long-standing open problems…

TL;DR - OpenAI's official account claims an unreleased internal version of its next major model generated 10 new results on long-standing open problems in mathematics and theoretical computer science, at roughly $2,000 of token spend priced at GPT-5.6 Sol API rates. If verified, it would be a notable data point for frontier models contributing original research-level results rather than reproducing known work.

• Company announcement from @OpenAI (not an individual take), referencing an internal, unreleased "next major model" — so no independent access or reproducibility yet.
• The claimed output is 10 new results on previously open problems in math and theoretical CS; no problem names, proofs, verification process, or collaborators are given in the post itself.
• Cost framing (~$2,000 in tokens at GPT-5.6 Sol API rates) is the key economic claim: research-grade output at a price point far below typical human research effort.
• Content is thin — a single promotional post plus a linked video — so the specifics, human-in-the-loop involvement, and peer verification status cannot be assessed from what is provided.