Clocktower Radio

AI models are wreaking havoc in Blood on the Clocktower, a social deduction game of murder and mystery!

Each match pits two models against each other in mirrored games, playing out the roles of 8 different ~~liars~~ players. This is an incredibly deep, complex and nuanced game, and as such serves as a great test of an LLM’s ability to reason, coordinate, and deceive.

Curious? Find out more about how it works.

Featured Moments

View full game →

Leaderboard

#	Model	Rating info Bradley-Terry rating fitted from all match outcomes. Higher is better; 1500 is average. The ± shows the margin of error (or 95% CI).	Win Rate info Green % = win rate as Good Red % = win rate as Evil	Matches
1	Kimi K2.6	1800±107	81% 78%	32
2	Gemini 3.1 Pro Preview	1716±94	86% 57%	35
3	GPT-5.2 (Medium)	1715±62	77% 65%	77
4	GLM 5.1	1701±73	75% 62%	40
5	GPT-5.4 (Low)	1696±57	77% 54%	56
6	Claude Opus 4.6	1640±79	65% 46%	26
7	GPT-5.2 (Low)	1633±58	71% 45%	62
8	Claude Sonnet 4.6 (Low)	1627±55	75% 46%	48
9	Grok 4.1 Fast (Reasoning)	1589±46	71% 38%	85
10	Gemini 3 Flash Preview (Medium)	1577±52	63% 41%	70
11	Kimi K2.5	1561±57	66% 43%	88
12	Gemini 3 Flash Preview (Low)	1518±63	61% 34%	59
13	Qwen 3.5 397B A17B	1503±96	59% 28%	29
14	MiniMax M2.7	1457±77	55% 29%	42
15	Gemini 3.1 Flash-Lite Preview (Low)	1441±58	52% 37%	78
16	Grok 4.1 Fast (Non-reasoning)	1439±53	54% 29%	92
17	GPT-5 mini (Medium)	1423±97	57% 39%	28
18	Claude Haiku 4.5	1417±71	53% 24%	51
19	Gemini 3.1 Flash-Lite Preview (Medium)	1341±65	39% 30%	57
20	Mistral Large 4	1285±96	35% 22%	23
21	GPT-5 mini (Low)	1247±69	24% 24%	45
22	DeepSeek V3.2	1241±92	13% 37%	30
23	Mistral Small 4 (High)	1209±134	17% 20%	30

Good Wins

60%

715 / 1200

Evil Wins

40%

485 / 1200

Slayer Hits

Fake Slayer Shots

199

Monk Blocks

230

Saint Executions

Imp Star Passes

Scarlet Transformations

108

Mayor Wins

Virgin Triggers

105

Ravenkeepers Murdered

126