Back to BenchmarksFrontier
Lech Mazur Writing
Writing quality and style evaluation
Leaderboard
(27 models)| Rank | Model | Score | Stderr |
|---|---|---|---|
| 1 | GPT-5.2 | 8.72 | — |
| 2 | Qwen3-Max-Instruct | 8.71 | — |
| 3 | kimi-k2-thinking (official) | 8.69 | — |
| 4 | o3 | 8.63 | — |
| 5 | Gemini 2.5 Pro (Jun 2025) | 8.60 | — |
| 6 | Claude Opus 4.5 | 8.54 | — |
| 7 | DeepSeek V3 | 8.52 | — |
| 8 | Qwen 3 235B | 8.49 | — |
| 9 | Qwen3-235B-A22B | 8.30 | — |
| 10 | DeepSeek R1 | 8.30 | — |
| 11 | GPT-4o | 8.18 | — |
| 12 | Grok 4 | 8.11 | — |
| 13 | Claude 3.7 Sonnet | 8.11 | — |
| 14 | QwQ-32B | 8.07 | — |
| 15 | Gemma 3 27B | 7.99 | — |
| 16 | GPT-OSS 120B | 7.73 | — |
| 17 | Grok-3 mini | 7.64 | — |
| 18 | GPT-4.1 | 7.56 | — |
| 19 | o4-mini-2025-04-16 medium | 7.50 | — |
| 20 | Gemini 2.0 Flash Thinking Exp | 7.38 | — |
| 21 | Claude 3.5 Haiku | 7.35 | — |
| 22 | Qwen2.5-Max | 7.29 | — |
| 23 | o1 | 7.02 | — |
| 24 | Mistral Large | 6.90 | — |
| 25 | gpt-4o-mini-2024-07-18 | 6.72 | — |
| 26 | Llama-4-Maverick-17B-128E-Instruct | 6.37 | — |
| 27 | Phi-4 | 6.26 | — |
Data source: Epoch AI, “Data on AI Benchmarking”. Published at epoch.ai
Licensed under CC-BY 4.0