CLEM Taboo — leaderboard
Metric: Game Clemscore (%). Source: huggingface.co. 31 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.2 (High) | 98.33 |
| 2 | GPT-5.2 (Medium) | 95.83 |
| 3 | GPT-5.2 | 95.56 |
| 4 | Gemini 3 Flash | 95.56 |
| 5 | Claude Sonnet 4.5 (High) | 95 |
| 6 | Claude Sonnet 4.5 | 94.16 |
| 7 | GLM-4.7 | 85.83 |
| 8 | Kimi K2 (Thinking) | 81.39 |
| 9 | DeepSeek V3.2 | 79.16 |
| 10 | GPT-OSS-120B | 77.77 |
| 11 | Mistral Large 3 | 76.67 |
| 12 | Qwen 3 Max | 76.67 |
| 13 | Kimi K2.5 (Non-reasoning) | 75.84 |
| 14 | GLM-5 (Non-reasoning) | 75.83 |
| 15 | MiniMax-M2.5 | 73.61 |
Interactive version: theaggregate.ai/benchmark?slug=clem-taboo · How the rankings work · Data refreshed daily, snapshot 2026-07-22.