Pencil Puzzle Bench - LITS — leaderboard
Metric: Direct-ask Success Rate (%). Source: ppbench.com. 51 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.2 (xHigh) | 53.3 |
| 2 | GPT-5.2 (Medium) | 40 |
| 3 | Claude Opus 4.6 (Thinking) | 40 |
| 4 | GPT-5.2 (High) | 33.3 |
| 5 | Kimi K2.5 | 13.3 |
| 6 | Claude Opus 4.5 (Thinking) | 13.3 |
| 7 | Claude Sonnet 4.6 (Thinking) | 13.3 |
| 8 | Gemini 3.1 Pro (Preview) | 6.7 |
| 9 | Gemini 3 Pro | 6.7 |
| 10 | GPT-5 (Medium) | 6.7 |
| 11 | Grok 4.1 Fast | 6.7 |
| 12 | GPT-5.1 (Medium) | 6.7 |
| 13 | GPT-5.2 Pro | 6.7 |
| 14 | Gemini 3 Flash (High) | 6.7 |
| 15 | Gemini 2.5 Pro | 0 |
Interactive version: theaggregate.ai/benchmark?slug=pencil-puzzle-bench-lits · How the rankings work · Data refreshed daily, snapshot 2026-07-22.