Dwarkesh Patel builds a Blood on the Clocktower benchmark testing AI deception
AIDwarkesh Patel says he built a benchmark around the social deception game Blood on the Clocktower, spending about $20,000 in API credits to run AI models against each other. He reports that GPT-6 Astra beats Claude Opus 4.6 by about 430 Elo points, winning roughly 92% of their games.
Loading source posts…
Source details
Dwarkesh Podcast · Dwarkesh Patel · dwarkesh.comPublished · added here
Original title: The benchmark where AIs have to lie to win