AtelierEval - Prompt Quality (Skilled Prompting): leaderboard

Metric: Judge score (1-5; mean of four prompt-level Likert dimensions from a memory-augmented GPT-5.4 judge calibrated on expert-scored exemplars, mean of the three task categories; structured prompting with chain-of-thought and self-verification). Source: arxiv.org. Saturation forecast: Around February 2027. 8 models tracked.

Top models

#ModelScore
1GPT-5.24.1
2Gemini 3 Pro (Preview)4.01
3Qwen 3 VL 235B A22B3.92
4Claude Sonnet 4.53.77
5GPT-4.13.72
6GPT-4.1 Nano3.48
7Qwen 3 VL 8B3.38
8Gemini 2.0 Flash3.23

Interactive version: theaggregate.ai/benchmark?slug=ateliereval-prompt-quality-skilled-prompting · How It Works · Data refreshed daily, snapshot 2026-09-26.