Assertion Supported AI assessment confidence: 95% certainty 4/5 debate potential 4/5

Brockman: Frontier AI models are solving unsolved math and physics problems

Greg Brockman · Big Technology AI Summit (full): Greg Brockman, Mike Krieger, Aaron Levie & Friends of The Podcast · Jun 26, 2026 · at 3:07:59

OpenAI President Greg Brockman describes the rapid capability gains of frontier AI systems.

0:00 / 0:06exact quote · 6.1s

▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark

“These models are now solving unsolved math problems and physics problems, right? Really helping people be able to achieve things they couldn't otherwise.”

quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →

The claim, plainly

Frontier AI models are currently solving previously unsolved problems in mathematics and physics.

a quote alone can be ambiguous, so this restates it with the references and context of the conversation around it resolved. AI-written; the tape below the player is the source.

Assessment

The claim asserted that frontier AI models were solving previously unsolved problems in both mathematics and physics by June 2026. Documented public evidence and peer-reviewed literature corroborate both components prior to the June 26, 2026 claim date. In mathematics, DeepMind's FunSearch (December 2023, Nature) found new extremal bounds, Epoch AI verified frontier models solving a 2019 open conjecture (March 2026), OpenAI models disproved Erdős's 1946 planar unit distance conjecture (May 2026, confirmed by independent mathematicians), and DeepMind's AlphaProof Nexus provided Lean-verified proofs for multiple open conjectures (May 2026). In theoretical physics, researchers from Google Research, Harvard, and CMU demonstrated that Gemini Deep Think derived closed-form exact analytical solutions for an open integral problem in cosmic string gravitational radiation (March 2026, arXiv:2603.04735).

Google DeepMind FunSearch published in Nature (December 14, 2023) discovering new bounds for the open cap set problem.
Brenner et al. (March 2026, arXiv:2603.04735), 'Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery', resolving an open integral problem in cosmic string theory via Gemini Deep Think.
OpenAI reasoning model disproving the 1946 Erdős planar unit distance conjecture in May 2026, verified by external mathematicians (arXiv:2605.20695).
DeepMind AlphaProof Nexus (May 2026, arXiv:2605.22763) generating Lean-verified proofs for 9 previously open Erdős problems and 44 open conjectures.

Checked against

[1] deepmind.google [2] wikipedia.org [3] substack.com [4] quantamagazine.org [5] livescience.com [6] theguardian.com [7] youtube.com [8] youtube.com

AI solved previously unsolved problem mathematics physics frontier AI model solves open math problem FunSearch open problem cap set Nature DeepMind AI solves open problem mathematics physics Gemini Deep Think cosmic strings physics open problem "Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery"

assessed as of 2026-09-02

This assessment is an AI reading of a public statement, made against the cited references with the confidence shown. It can be wrong. The quote, the tape and the references are all on this page; if something is off, use report an error below. full disclaimer →

Mentioned

OpenAI

⚑ Spot an error? Report it

More from Greg Brockman

Prediction Not checkable as stated

Brockman: OpenAI Has Definitive Line of Sight to AGI in 2026

“And I think that we have definitively answered that question of it's, it is going to go to AGI. Like we see line of sight and that it is at this point, we have line of sight to these much better models that are coming this year.”

Greg Brockman Apr 1, 2026 ▶ 7:44 OpenAI President Greg Brockman: AI Self-Improvement, The Superapp Bet, Path To AGI, Scaling Compute

Prediction Not checkable as stated

Brockman: AGI Capable of Any Computer Task Will Arrive Within Years

“I think it's extremely clear that we are going to have AGI within the next couple years in a way that is still going to be jagged, but that the floor of task will just be almost for any intellectual task of how you use your computer. The AI will be able to do …”

Greg Brockman Apr 1, 2026 ▶ 37:25 OpenAI President Greg Brockman: AI Self-Improvement, The Superapp Bet, Path To AGI, Scaling Compute

Assertion Supported

Brockman: AI models are now solving unsolved math and physics problems

“These models are now solving unsolved math problems and physics problems, right?”

Greg Brockman Jul 1, 2026 ▶ 5:45 OpenAI President Greg Brockman: Our Plan To Merge Chat And Agents

Opinion

Brockman: No layer of the AI stack will lose value capture

“Well, look, I don't think that there's any layer of the stack here that is going to just kind of be removed from the value chain. I think that these things multiplied together.”

Greg Brockman Jul 1, 2026 ▶ 34:47 OpenAI President Greg Brockman: Our Plan To Merge Chat And Agents

Assertion Not checkable as stated

Brockman: Competitors Scrambled Too Late for Compute When None Remained Available

“I think that what we have seen is that for other players, that they kind of realized that probably late last year and started scrambling to see what compute is available, and there really wasn't any”

Greg Brockman Apr 1, 2026 ▶ 59:52 OpenAI President Greg Brockman: AI Self-Improvement, The Superapp Bet, Path To AGI, Scaling Compute

Disclosure

Brockman Confirms Contribution to MAGA Inc. Alongside Bipartisan PAC Donations

“So the way I look at this, so my wife and I made that donation. We've donated to bipartisan super PACs as well.”

Greg Brockman Apr 1, 2026 ▶ 1:09:24 OpenAI President Greg Brockman: AI Self-Improvement, The Superapp Bet, Path To AGI, Scaling Compute

Brockman: Frontier AI models are solving unsolved math and physics problems

More from Greg Brockman

Turn any episode into a week of clips.