Divergent Thinking: leaderboard
Creativity benchmark where LLMs generate 25 maximally distinct words unrelated to an initial 50-word list. Each word pair is judged by 4 LLMs. Tests originality and fluency.
Metric: Divergence Score. Source: github.com. Status: saturated. 19 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O1 Preview | 4.79 |
| 2 | Claude 3 Opus | 4.47 |
| 3 | Grok 2 (1212) | 4.45 |
| 4 | Llama 3.3 70B | 4.44 |
| 5 | Claude 3.5 Sonnet (20241022) | 4.41 |
| 6 | Gemini 2.0 Flash (Thinking) | 4.41 |
| 7 | Gemma 2 27B | 4.37 |
| 8 | O1 Mini | 4.2 |
| 9 | Claude 3.5 Haiku | 4.16 |
| 10 | Mistral Large 2 (Jul) | 4.14 |
| 11 | GPT-4o Mini | 4.12 |
| 12 | Gemini 1.5 Flash | 4.09 |
| 13 | Gemini 1.5 Pro (Sept) | 4.07 |
| 14 | Claude 3 Haiku | 3.98 |
| 15 | Qwen 2.5 72B | 3.89 |
Interactive version: theaggregate.ai/benchmark?slug=divergent-thinking · How It Works · Data refreshed daily, snapshot 2026-09-05.