HardcoreLogic - Unsolvable Puzzles — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 Mini | 98.52 |
| 2 | GPT-5 | 97.78 |
| 3 | Grok 4 | 97.59 |
| 4 | DeepSeek R1 0528 Qwen3 8B | 95.19 |
| 5 | O4 Mini | 95 |
| 6 | Grok 3 Mini | 94.63 |
| 7 | DeepSeek R1 0528 | 93.5 |
| 8 | GPT-OSS-120B | 93.35 |
| 9 | GLM-4.5 | 93.26 |
| 10 | Gemini 2.5 Pro | 91.48 |
| 11 | DeepSeek V3.1 | 88.76 |
| 12 | Kimi K2 | 87.46 |
| 13 | Qwen 3 30B A3B 2507 (Thinking) | 86.09 |
| 14 | Seed-OSS-36B-Instruct | 85.76 |
| 15 | Qwen 3 235B A22B (Thinking) | 84.41 |
Interactive version: aibenchmarks.dev/benchmark?slug=hardcorelogic-unsolvable-puzzles · How the rankings work · Data refreshed daily, snapshot 2026-07-17.