EQ-Bench Creative Writing v3 — leaderboard
LLM-judged creative writing benchmark measuring abilities, style, repetitiveness, and 'slop' (overused AI-typical phrases). Includes vocabulary and GPT-slop control metrics.
Metric: Elo. Source: eqbench.com. Status: saturation imminent. 106 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Fable 5 | 2027.9 |
| 2 | Claude Opus 4.7 | 2015.8 |
| 3 | GPT-5.5 | 1861.3 |
| 4 | Claude Opus 4.8 | 1831.1 |
| 5 | GPT-5.4 | 1830.3 |
| 6 | Claude Sonnet 4.6 | 1790.5 |
| 7 | Claude Opus 4.6 | 1769.9 |
| 8 | GLM-5.2 | 1643.2 |
| 9 | GPT-5.2 | 1628.4 |
| 10 | Kimi K2.6 | 1618.7 |
| 11 | Claude Sonnet 4.5 | 1607.8 |
| 12 | Claude Opus 4.5 (20251101) | 1598.2 |
| 13 | O3 | 1584.2 |
| 14 | Kimi K2 | 1580.1 |
| 15 | GPT-5.4 Mini | 1566.7 |
Interactive version: theaggregate.ai/benchmark?slug=eq-bench-creative-writing-v3 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.