AI BENCHY Category Failures
Puzzle Solving: Wrong answer
Puzzle Solving
Wrong answer
See which AI models are most likely to hit Wrong answer on Puzzle Solving, so you can spot weak points faster. Sort by: Tests Correct ↑.
Failure Reasons
| Rank | Model | Company | Wrong answer Count | Category Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #70 | GPT-5.4 Nano medium | OpenAI | 2 | 4.1 | 0/3 | 3.79s |
| #105 | Nemotron 3 Super medium | NVIDIA | 2 | 3.0 | 0/3 | 3.15s |
| #108 | Qwen3.5-Flash none | Qwen | 3 | 3.1 | 0/3 | 10.9s |
| #117 | Qwen3.5-35B-A3B none | Qwen | 2 | 3.7 | 0/3 | 1.35s |
| #119 | Cobuddy medium | Baidu | 2 | 3.6 | 0/3 | 12.8s |
| #124 | Kimi K2.6 none | Moonshot AI | 2 | 3.1 | 0/3 | 1.40s |
| #128 | Qwen3.6 Flash none | Qwen | 2 | 3.5 | 0/3 | 1.21s |
| #131 | Qwen3.5-122B-A10B none | Qwen | 2 | 3.8 | 0/3 | 1.00s |
| #132 | Mistral Small 4 medium | Mistral | 2 | 3.4 | 0/3 | 2.17s |
| #135 | Kimi K2.5 none | Moonshot AI | 3 | 3.0 | 0/3 | 4.04s |
| #137 | Elephant Alpha none | Openrouter | 2 | 4.2 | 0/3 | 807ms |
| #138 | Ling-2.6-flash none | Inclusionai | 2 | 2.9 | 0/3 | 6.51s |
| #139 | DeepSeek V4 Flash none | DeepSeek | 1 | 3.1 | 0/3 | 23.7s |
| #140 | Qwen3 Coder Next none | Qwen | 3 | 3.0 | 0/3 | 24.3s |
| #142 | Mistral Small 4 none | Mistral | 2 | 3.1 | 0/3 | 399ms |