Clocktower Bench - Detection: leaderboard

Clocktower Bench (Dwarkesh Patel, 2026) evaluates strategic social deduction in Blood on the Clocktower: one model plays all good seats, another all evil seats, with isolated seat contexts and side-swapped deals. The published 386-game study pools the authors' v16/v17 protocols over 11 models, each at high effort and an 8,000-token per-turn cap, and reports detection Bradley–Terry Elo, centered at 1500. Detection measures good-town play; deception measures evil-team play. Ratings use the authors' setup adjustments and 90% side-swapped-pair bootstrap intervals. Retry-bug originals and mixed-team ablations are excluded from this fit. Higher is better; Elo has no theoretical bounds.

Metric: Native Bradley–Terry Elo rating for town-side detection; 1500 centers the rated model population; higher is better; unbounded. Source: clocktowerbench.com. Saturation forecast: Rough model projection: around 2026. 11 models tracked.

Top models

#ModelScoreOverall rank
1GPT-6 Astra (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1940
2GPT-6.1 Sol (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1793
3GPT-5.6 Sol (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1666
4Claude Fable 5.1 (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1582
5Claude Opus 5.5 (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1500
6Claude Opus 4.6 (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1489
7Claude Fable 5 (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1438
8GPT-5.5 (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1380
9GPT-5.4 (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1309
10Claude Opus 4.7 (Clocktower Run: study=2026-v16-v17; effort=high; max_tokens=8000)1266

Interactive version: theaggregate.ai/benchmark?slug=clocktower-bench-detection · How It Works · Data refreshed daily, snapshot 2026-11-06.