Creative Writing (Lechmazur): leaderboard

Evaluates creative fiction writing quality across 30+ models using structured prompts and LLM-as-judge scoring. Measures prose quality, creativity, coherence, and adherence to constraints.

Metric: Mean Score. Source: github.com. Status: years away from saturation. 48 models tracked.

Top models

#ModelScore
1Claude Opus 5 (xHigh)4.1
2Claude Fable 5.1 (High)4.1
3GLM-5.3 (Max)3.4
4Claude Fable 5 (High)3.1
5Kimi K32.8
6GPT-5.5 (xHigh)2.7
7GPT-5.6 Sol (xHigh)2.7
8GPT-5.6 Sol (High)2.6
9GPT-5.4 (xHigh)2.2
10GPT-5.4 (Medium)2.1
11Claude Opus 4.72
12Claude Opus 4.8 (xHigh)1.1
13Muse Spark 1.1 (High)1.1
14DeepSeek V4 Pro (High)0.8
15GLM-5.2 (Max)0.7

Interactive version: theaggregate.ai/benchmark?slug=creative-writing-lechmazur · How It Works · Data refreshed daily, snapshot 2026-09-05.