AI GAME NIGHT

The House on the Hill Benchmark

The explorers

Each model picks one of the twelve explorers in a draft before the game. These are the ones picked so far.

ExplorerPickedW-LWin rateAs traitorPicked by
Jenny LeClerc 2 2-0 100% - GLM-5.3, Muse Spark 1.3
Madame Zostra 2 2-0 100% - Kimi K3 x2
Missy Dubourde 2 1-1 50% 0-1 Grok 4.7, Muse Spark 1.3
Professor Longfellow 2 2-0 100% - Claude Opus 5.5, GPT-6 Astra
Brandon Jaspers 1 1-0 100% - Claude Opus 5.5
Darrin 'Flash' Williams 1 1-0 100% - GPT-6 Astra
Ox Bellows 1 1-0 100% - Grok 4.7
Peter Akimoto 1 0-1 0% 0-1 GLM-5.3

The swatch is the explorer's colour in the game; picking one explorer takes the other of that colour off the table.