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. Saturation forecast: Estimated already saturated. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O1 Preview | 4.79 | #193 |
| 2 | Gemini 2.0 Flash (Preview) | 4.65 | #370 |
| 3 | Claude 3 Opus | 4.47 | #571 |
| 4 | Grok 2 (1212) | 4.45 | #496 |
| 5 | Llama 3.3 70B | 4.44 | #579 |
| 6 | Claude 3.5 Sonnet (20241022) | 4.41 | #291 |
| 7 | Gemini 2.0 Flash (Thinking) | 4.41 | #342 (Gemini 2.0 Flash) |
| 8 | Gemma 2 27B | 4.37 | #806 |
| 9 | O1 Mini | 4.2 | #323 |
| 10 | Claude 3.5 Haiku | 4.16 | #568 |
| 11 | Mistral Large 2 (Jul) | 4.14 | #484 |
| 12 | GPT-4o Mini | 4.12 | #599 |
| 13 | Gemini 1.5 Flash | 4.09 | #618 |
| 14 | Gemini 1.5 Pro (Sept) | 4.07 | #448 |
| 15 | Claude 3 Haiku | 3.98 | #803 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=divergent-thinking · How It Works · Data refreshed daily, snapshot 2026-10-20.