Puzzle Solving Ranking
See which AI models perform best on Puzzle Solving, which ones stay reliable, and where the biggest gaps appear. Sort by: Total Cost ↓.
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) |
|---|---|---|---|---|---|---|---|
| #84 | DeepSeek V4 Flash 0423 high | DeepSeek | 8.2 | 7.6 | $0.039 | 2/3 | 26.1s |
| #233 | DeepSeek V4 Flash 0423 none | DeepSeek | 3.1 | 5.4 | $0.037 | 0/3 | 23.7s |
| #283 | Qwen3 Coder Next medium | Qwen | 3.0 | 4.6 | $0.034 | 0/3 | 1.25s |
| #279 | Laguna M.1 medium | Poolside | 5.3 | 4.7 | $0.033 | 1/3 | 10.2s |
| #281 | Mercury 2 none | Inception | 3.1 | 4.7 | $0.030 | 0/3 | 535ms |
| #163 | Mercury 2.5 Preview high | Inception | 10.0 | 6.6 | $0.030 | 3/3 | 1.21s |
| #236 | GPT-5.6 Luna none | OpenAI | 5.3 | 5.4 | $0.029 | 1/3 | 790ms |
| #111 | GLM 5.3 Flash high | Z.ai | 10.0 | 7.3 | $0.029 | 3/3 | 3.04s |
| #213 | GLM 5 none | Z.ai | 7.7 | 5.7 | $0.027 | 2/3 | 1.91s |
| #269 | Ring-2.6-1T none | Inclusionai | 7.7 | 4.8 | $0.026 | 2/3 | 31.5s |
| #254 | Qwen3 Coder Next none | Qwen | 3.0 | 5.1 | $0.026 | 0/3 | 24.3s |
| #291 | Granite 4.2 8B none | IBM Granite | 3.0 | 4.3 | $0.026 | 0/3 | 107.6s |
| #255 | MiMo-V2.5 none | Xiaomi | 5.4 | 5.1 | $0.025 | 1/3 | 2.13s |
| #296 | MiMo-V2-Flash none | Xiaomi | 5.3 | 4.0 | $0.025 | 1/3 | 1.86s |
| #96 | Mercury 2.5 Preview medium | Inception | 10.0 | 7.5 | $0.024 | 3/3 | 1.87s |