Creative Story-Writing (Lech Mazur)
Short stories written to the same constrained briefs, compared head to head by a panel of evaluator models; the score is the average predicted chance of beating another writer model.
| Category | General |
|---|---|
| Metric | Win chance (%) |
| Source | Lech Mazur (original) |
| License | None stated |
| Benchmark page | github.com |
| Status | Scored |
| Weight | 0.8 within General |
| Fit | Discrimination 1.88, noise 0.67 |
Results
| # | Model | Provider | Result | Date |
|---|---|---|---|---|
| 1 | Claude Opus 5.5 High | Anthropic | 93.0% | 2026-10-23 |
| 2 | Claude Fable 5.1 High | Anthropic | 93.0% | 2026-10-23 |
| 3 | Claude Opus 5 Extra High | Anthropic | 93.0% | 2026-10-23 |
| 4 | GPT-6 Astra High | OpenAI | 91.0% | 2026-10-23 |
| 5 | Claude Opus 5 High | Anthropic | 90.0% | 2026-10-23 |
| 6 | GLM-5.3 Max | Zhipu AI | 87.0% | 2026-10-23 |
| 7 | Claude Fable 5 High | Anthropic | 86.0% | 2026-10-23 |
| 8 | Kimi K3.1 Pro High | Moonshot AI | 85.0% | 2026-10-23 |
| 9 | GPT-5.5 Extra High | OpenAI | 82.0% | 2026-10-23 |
| 10 | GPT-5.6 Sol Extra High | OpenAI | 82.0% | 2026-10-23 |
| 11 | Kimi K3 | Moonshot AI | 81.0% | 2026-10-23 |
| 12 | GPT-5.6 Sol High | OpenAI | 80.0% | 2026-10-23 |
| 13 | GPT-5.4 Medium | OpenAI | 77.0% | 2026-10-23 |
| 14 | GPT-5.4 Extra High | OpenAI | 76.0% | 2026-10-23 |
| 15 | Muse Wonder High | Meta | 76.0% | 2026-10-23 |
| 16 | Claude Opus 4.7 High | Anthropic | 74.0% | 2026-10-23 |
| 17 | Claude Sonnet 4.6 Thinking | Anthropic | 73.0% | 2026-10-23 |
| 18 | MiMo-V2.6-Pro Thinking | Xiaomi | 72.0% | 2026-10-23 |
| 19 | Claude Opus 4.6 Thinking | Anthropic | 66.0% | 2026-10-23 |
| 20 | Gemini 4 Argon High | 61.0% | 2026-10-23 | |
| 21 | Claude Opus 4.8 Extra High | Anthropic | 61.0% | 2026-10-23 |
| 22 | Muse Spark 1.1 High | Meta | 61.0% | 2026-10-23 |
| 23 | Muse Spark 1.3 High | Meta | 57.0% | 2026-10-23 |
| 24 | DeepSeek V4 Pro High | DeepSeek | 57.0% | 2026-10-23 |
| 25 | GLM-5.2 Max | Zhipu AI | 56.0% | 2026-10-23 |
| 26 | Grok 4.7 High | xAI | 55.0% | 2026-10-23 |
| 27 | GPT-5.2 Medium | OpenAI | 54.0% | 2026-10-23 |
| 28 | Claude Opus 4.8 High | Anthropic | 53.0% | 2026-10-23 |
| 29 | Qwen3.8 Max | Alibaba | 52.0% | 2026-10-23 |
| 30 | Gemini 3.8 Flash High | 51.0% | 2026-10-23 | |
| 31 | Kimi K2.6 | Moonshot AI | 50.0% | 2026-10-23 |
| 32 | MiniMax-M3 | MiniMax | 48.0% | 2026-10-23 |
| 33 | Muse Spark 1.2 High | Meta | 48.0% | 2026-10-23 |
| 34 | Mistral Medium 3.1 | Mistral | 43.0% | 2026-10-23 |
| 35 | DeepSeek V4.1 Flash High | DeepSeek | 42.0% | 2026-10-23 |
| 36 | MiMo-V2.5-Pro | Xiaomi | 39.0% | 2026-10-23 |
| 37 | Qwen3.8 27B | Alibaba | 39.0% | 2026-10-23 |
| 38 | Qwen3 Max Preview | Alibaba | 39.0% | 2026-10-23 |
| 39 | Gemini 3.7 Flash High | 39.0% | 2026-10-23 | |
| 40 | Qwen3.6 Max Preview | Alibaba | 35.0% | 2026-10-23 |
| 41 | GLM-5.1 | Zhipu AI | 34.0% | 2026-10-23 |
| 42 | Kimi K2.5 Thinking | Moonshot AI | 33.0% | 2026-10-23 |
| 43 | MiMo-V2-Pro | Xiaomi | 31.0% | 2026-10-23 |
| 44 | Mistral Large 4 High | Mistral | 26.0% | 2026-10-23 |
| 45 | Mistral Large 3 | Mistral | 24.0% | 2026-10-23 |
| 46 | Gemma 4 31B IT Thinking | 23.0% | 2026-10-23 | |
| 47 | Gemini 3.5 Flash | 22.0% | 2026-10-23 | |
| 48 | Seed 2.0 Pro | ByteDance | 22.0% | 2026-10-23 |
| 49 | Gemini 3.1 Pro Preview | 19.0% | 2026-10-23 | |
| 50 | Qwen3.6 Plus | Alibaba | 18.0% | 2026-10-23 |
| 51 | Mistral Medium 3.5 | Mistral | 16.0% | 2026-10-23 |
| 52 | Qwen3.7 Max | Alibaba | 14.0% | 2026-10-23 |
| 53 | DeepSeek V3.2 | DeepSeek | 13.0% | 2026-10-23 |
| 54 | Grok 4.6 High | xAI | 12.0% | 2026-10-23 |
| 55 | GPT OSS 120B | OpenAI | 11.0% | 2026-10-23 |
| 56 | MiniMax-M2.7 | MiniMax | 6.0% | 2026-10-23 |
| 57 | Grok 4.3 | xAI | 4.0% | 2026-10-23 |
| 58 | Grok 4.5 High | xAI | 2.0% | 2026-10-23 |