OpenAI Claims Frontier Models Refuted Fundamental Exponential Time Hypothesis

27 Oct 2026, 10:38 pm · 3h ago · 1 min read · TechCrunch AI

Frontier artificial intelligence laboratories, including OpenAI, claim their advanced reasoning models have generated counterexamples refuting the Exponential Time Hypothesis, a foundational conjecture in theoretical computer science. The hypothesis suggests certain computational problems inherently require exponential time to solve. In discussions with Massachusetts Institute of Technology researchers, computer scientists are analyzing whether these machine-generated proofs represent genuine mathematical breakthroughs or subtle algorithmic hallucinations. The discovery demonstrates the expanding role of advanced artificial intelligence models in exploring complex computational complexity theory and challenging longstanding academic assumptions about algorithm limits, though formal peer verification remains underway.