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 ↑.
311/311
Filter models
No models match the current search and filters.
| Rank | Model | Company | Puzzle Solving Score | Score | Total Cost | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|---|
| #50 | Inkling Small high | Thinkingmachines | 4.8 | 8.4 | $0.398 | 0/3 | 14.4s |
| #90 | GPT-5.4 Nano medium | OpenAI | 4.1 | 7.5 | $0.142 | 0/3 | 3.79s |
| #152 | LongCat 2.0 low | Meituan | 3.1 | 6.7 | $0.412 | 0/3 | 8.15s |
| #155 | LongCat 2.0 high | Meituan | 3.1 | 6.7 | $0.492 | 0/3 | 9.18s |
| #172 | LongCat 2.0 none | Meituan | 4.0 | 6.4 | $0.044 | 0/3 | 2.74s |
| #185 | Qwen3.7 Flash none | Qwen | 3.7 | 6.1 | $0.019 | 0/3 | 1.39s |
| #189 | Trinity Large Thinking medium | Arcee AI | 5.2 | 6.1 | $0.756 | 0/3 | 2.90s |
| #190 | Qwen3.6 Flash none | Qwen | 3.5 | 6.1 | $0.062 | 0/3 | 1.21s |
| #200 | Qwen3.5-Flash none | Qwen | 3.1 | 6.0 | $0.073 | 0/3 | 10.9s |
| #201 | Qwen3.5-35B-A3B none | Qwen | 3.7 | 5.9 | $0.142 | 0/3 | 1.35s |
| #211 | North Mini Code medium | Cohere | 3.3 | 5.7 | $0.000 | 0/3 | 19.7s |
| #212 | Nemotron 3 Super medium | NVIDIA | 3.0 | 5.7 | $0.050 | 0/3 | 3.15s |
| #217 | Kimi K2.6 none | Moonshot AI | 3.1 | 5.7 | $0.233 | 0/3 | 1.40s |
| #218 | Qwen3.5-122B-A10B none | Qwen | 3.8 | 5.7 | $0.283 | 0/3 | 1.00s |
| #220 | KAT-Coder-Air V2.5 high | Kwaipilot | 3.5 | 5.6 | $0.077 | 0/3 | 2.47s |