OpenAI insiders hint at huge math capability overhang: models could be far better but aren't a focus

pwlot · x · 2026-09-07

Jeff Ladish relays that OpenAI's merettm said the lab could make its models much better at math but hasn't because it's not a focus — implying the recent unprecedented math progress could go further. pwlot adds there are "lots and lots of overhang."

Related event: OpenAI Reportedly Held Back Its Models' Math Capabilities(2 posts)→

Original post →

More from Models

Models channel →