“I'd bet big against a Chinese lab actually having the best model at any point in 2026, but benchmarks are not leaderboards. I'd be very surprised if this happened on Artificial Analysis... I'd be less surprised if this happened on LM Arena, since it is rather hackable... I still think this is an underdog. [On the broader 'a Chinese lab tops a global leaderboard' wording: 'I'll give them maybe 55%... Manifold says 34%, and on reflection yeah I was being a coward and moderating my instincts too much, that's more like it.']”
Said Oct 10, 2025Deadline Dec 31, 2026Don't Worry About the Vase, '2025 State of AI Report and Predictions', October 10, 2025
Written in response to the State of AI Report 2025's prediction that 'A Chinese lab tops a global leaderboard'; resolvable at the end of 2026 against Artificial Analysis and other leaderboards.