Slop Score
Chat & Writing · 2025-10-25
Slop Score measures how often generated prose uses stock phrases and other stylistic habits associated with LLM writing. Lower scores indicate less formulaic prose.
Top models (lower is better)
| Model | Score |
|---|---|
| Kimi K2 Instruct 0905 | 18.3 |
| Sonnet 4.5 | 19.5 |
| Kimi K2 Instruct | 23.7 |
| GPT-5 Mini | 26.3 |
| Sonnet 4 | 27.3 |
| o3 | 31.8 |
| GPT-5 Nano | 34.1 |
| ChatGPT-4o Latest (source-unspecified snapshot) | 47.7 |
| Llama 4 Maverick Instruct | 49.2 |
| Mistral Small 3.2 24B Instruct 2506 | 52.7 |
| DeepSeek-V3.2-Exp | 54.2 |
| Mistral NeMo Instruct 2407 | 54.4 |
| Ling-1T | 56.2 |
| GLM-4.5 | 59.9 |
| Qwen3 4B | 62.2 |
| DeepSeek-R1 | 67.9 |
| Gemma 3 27B IT | 69.5 |
| Gemma 3 12B IT | 70.6 |
| Gemini 2.5 Flash | 77.6 |
| Gemma 3 4B IT | 85.2 |