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.
| Rank | Model | Company | Puzzle Solving Score | Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #27 | Gemma 4 31B medium | 9.9 | 7.8 | 3/3 | 26.9s | |
| #15 | GPT-5.3-Codex medium | OpenAI | 9.0 | 8.4 | 2/3 | 5.05s |
| #19 | Seed-2.0-Lite medium | Bytedance Seed | 9.0 | 8.2 | 2/3 | 10.2s |
| #23 | GLM 5 Turbo medium | Z.ai | 8.7 | 8.0 | 2/3 | 5.23s |
| #21 | GPT-5.4 medium | OpenAI | 8.2 | 8.0 | 2/3 | 9.14s |
| #31 | DeepSeek V4 Flash high | DeepSeek | 8.2 | 7.7 | 2/3 | 26.1s |
| #36 | Qwen3.5 Plus 2026-04-20 medium | Qwen | 8.2 | 7.6 | 2/3 | 17.7s |
| #39 | Qwen3.6 Flash medium | Qwen | 8.2 | 7.5 | 2/3 | 6.29s |
| #49 | Qwen3.5-Flash medium | Qwen | 8.2 | 7.4 | 2/3 | 27.6s |
| #55 | GLM 5.1 medium | Z.ai | 8.2 | 7.3 | 2/3 | 31.6s |
| #56 | MiMo-V2.5 medium | Xiaomi | 8.2 | 7.3 | 2/3 | 20.3s |
| #66 | Qwen3.5-35B-A3B medium | Qwen | 8.2 | 7.1 | 2/3 | 33.1s |
| #73 | Seed-2.0-Mini medium | Bytedance Seed | 8.2 | 6.9 | 2/3 | 31.8s |
| #30 | Qwen3.5-27B medium | Qwen | 8.2 | 7.8 | 2/3 | 59.6s |
| #46 | Qwen3.6 35B A3B medium | Qwen | 8.0 | 7.4 | 2/3 | 5.95s |