AI BENCHY Category
Puzzle Solving Ranking
See which AI models perform best on Puzzle Solving, which ones stay reliable, and where the biggest gaps appear. Sort by: Tests Correct ↑.
| Rank | Model | Company | Puzzle Solving Score | Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #70 | GPT-5.4 Nano medium | OpenAI | 4.1 | 7.0 | 0/3 | 3.79s |
| #105 | Nemotron 3 Super medium | NVIDIA | 3.0 | 5.8 | 0/3 | 3.15s |
| #108 | Qwen3.5-Flash none | Qwen | 3.1 | 5.8 | 0/3 | 10.9s |
| #117 | Qwen3.5-35B-A3B none | Qwen | 3.7 | 5.6 | 0/3 | 1.35s |
| #119 | Cobuddy medium | Baidu | 3.6 | 5.6 | 0/3 | 12.8s |
| #124 | Kimi K2.6 none | Moonshot AI | 3.1 | 5.5 | 0/3 | 1.40s |
| #128 | Qwen3.6 Flash none | Qwen | 3.5 | 5.4 | 0/3 | 1.21s |
| #131 | Qwen3.5-122B-A10B none | Qwen | 3.8 | 5.3 | 0/3 | 1.00s |
| #132 | Mistral Small 4 medium | Mistral | 3.4 | 5.3 | 0/3 | 2.17s |
| #135 | Kimi K2.5 none | Moonshot AI | 3.0 | 5.2 | 0/3 | 4.04s |
| #137 | Elephant Alpha none | Openrouter | 4.2 | 5.1 | 0/3 | 807ms |
| #138 | Ling-2.6-flash none | Inclusionai | 2.9 | 5.0 | 0/3 | 6.51s |
| #139 | DeepSeek V4 Flash none | DeepSeek | 3.1 | 5.0 | 0/3 | 23.7s |
| #140 | Qwen3 Coder Next none | Qwen | 3.0 | 4.9 | 0/3 | 24.3s |
| #142 | Mistral Small 4 none | Mistral | 3.1 | 4.9 | 0/3 | 399ms |