Pencil Puzzle Bench - Tapa — leaderboard

Metric: Direct-ask Success Rate (%). Source: ppbench.com. 51 models tracked.

Top models

#ModelScore
1Claude Opus 4.6 (Thinking)60
2GPT-5.2 (xHigh)60
3Gemini 3.1 Pro (Preview)40
4GPT-5.2 (High)33.3
5Kimi K2.520
6GPT-5.2 Pro20
7Claude Sonnet 4.6 (Thinking)20
8Grok 4.1 Fast (Reasoning)13.3
9O313.3
10GPT-5.2 (Medium)13.3
11Gemini 3 Flash (High)13.3
12Gemini 2.5 Pro6.7
13Claude Sonnet 4.66.7
14GPT-5.26.7
15GPT-OSS-120B6.7

Interactive version: theaggregate.ai/benchmark?slug=pencil-puzzle-bench-tapa · How the rankings work · Data refreshed daily, snapshot 2026-07-22.