CLEM Wordle with Critic — leaderboard
Metric: Game Clemscore (%). Source: huggingface.co. 30 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.5 (High) | 86.11 |
| 2 | Claude Sonnet 4.5 | 76.39 |
| 3 | Gemini 3 Flash | 73.89 |
| 4 | GPT-5.2 (High) | 70.28 |
| 5 | DeepSeek V3.2 | 68.33 |
| 6 | GPT-5.2 (Medium) | 66.67 |
| 7 | GPT-5.2 | 63.06 |
| 8 | Kimi K2 (Thinking) | 54.83 |
| 9 | Qwen 3 Max | 53.89 |
| 10 | GPT-OSS-120B | 50.39 |
| 11 | DeepSeek V3.2 (Non-reasoning) | 49.72 |
| 12 | GLM-5 (Non-reasoning) | 43.33 |
| 13 | Kimi K2.5 (Non-reasoning) | 38.34 |
| 14 | Qwen 3 Next 80B A3B Instruct | 36.11 |
| 15 | GLM-4.7 | 35 |
Interactive version: theaggregate.ai/benchmark?slug=clem-wordle-with-critic · How the rankings work · Data refreshed daily, snapshot 2026-07-22.