NonoBench: leaderboard
Nonogram puzzle solving benchmark testing LLM spatial reasoning and constraint satisfaction through Japanese logic puzzles.
Metric: Standard accuracy (%), correct / non-failed attempts; version 1.2. Source: nonobench.com. Saturation forecast: Around January 2027. 84 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra (xHigh) | 100 | #6 (GPT-6 Astra) |
| 2 | GPT-6 Astra (High) | 96.67 | #6 (GPT-6 Astra) |
| 3 | GPT-6 Astra (Max) | 93.33 | #6 (GPT-6 Astra) |
| 4 | GPT-6.1 Sol (Max) | 93.33 | #7 (GPT-6.1 Sol) |
| 5 | Claude Opus 5.5 (High) | 93.33 | #3 (Claude Opus 5.5) |
| 6 | GPT-6 Astra (Medium) | 90 | #6 (GPT-6 Astra) |
| 7 | GPT-6.1 Sol (High) | 90 | #7 (GPT-6.1 Sol) |
| 8 | GPT-6.1 Sol (Medium) | 90 | #7 (GPT-6.1 Sol) |
| 9 | Claude Opus 5.5 (Medium) | 90 | #3 (Claude Opus 5.5) |
| 10 | GPT-5.6 Sol (Max) | 86.67 | #19 (GPT-5.6 Sol) |
| 11 | GPT-5.6 Sol (xHigh) | 86.67 | #19 (GPT-5.6 Sol) |
| 12 | GPT-5.6 Sol (High) | 86.67 | #19 (GPT-5.6 Sol) |
| 13 | Gemini 4 Argon (High) | 86.67 | #8 (Gemini 4 Argon) |
| 14 | GPT-6 Astra (Low) | 86.67 | #6 (GPT-6 Astra) |
| 15 | GPT-6.1 Sol (xHigh) | 86.67 | #7 (GPT-6.1 Sol) |
No result here: #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=nonobench · How It Works · Data refreshed daily, snapshot 2026-10-20.