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: Response Time (avg) ↓.
| Rank | Model | Company | Puzzle Solving Score | Score | Tests Correct | Response Time (avg) |
|---|---|---|---|---|---|---|
| #111 | Owl Alpha medium | Openrouter | 5.3 | 5.7 | 1/3 | 3.40s |
| #2 | Gemini 3.5 Flash high | 10.0 | 9.6 | 3/3 | 3.23s | |
| #28 | Gemini 2.5 Flash medium | 7.7 | 7.8 | 2/3 | 3.18s | |
| #105 | Nemotron 3 Super medium | NVIDIA | 3.0 | 5.8 | 0/3 | 3.15s |
| #20 | Gemini 3.5 Flash none | 10.0 | 8.1 | 3/3 | 3.13s | |
| #63 | GPT-5.3 Chat none | OpenAI | 10.0 | 7.2 | 3/3 | 2.99s |
| #110 | Seed-2.0-Lite none | Bytedance Seed | 5.3 | 5.8 | 1/3 | 2.78s |
| #68 | Claude Opus 4.8 none | Anthropic | 7.7 | 7.0 | 2/3 | 2.74s |
| #95 | Qwen3.5 Plus 2026-02-15 none | Qwen | 7.7 | 6.3 | 2/3 | 2.71s |
| #74 | Qwen3.6 Max Preview none | Qwen | 10.0 | 6.9 | 3/3 | 2.65s |
| #134 | GLM 5 Turbo none | Z.ai | 5.5 | 5.2 | 1/3 | 2.65s |
| #77 | Claude Sonnet 4.6 none | Anthropic | 7.7 | 6.8 | 2/3 | 2.53s |
| #8 | Claude Opus 4.7 none | Anthropic | 10.0 | 8.9 | 3/3 | 2.46s |
| #11 | Claude Opus 4.7 medium | Anthropic | 10.0 | 8.7 | 3/3 | 2.43s |
| #109 | GLM 5V Turbo none | Z.ai | 5.3 | 5.8 | 1/3 | 2.40s |