AtelierEval - Image Constraint Satisfaction (Novice Prompting): leaderboard
Metric: Image checklist satisfaction (%; share of each task's expert constraint checklist realized in the generated image, best of four images per backend, mean over four text-to-image backends (Gemini 3 Pro Image, GPT-Image-1, Flux.1 Pro, SDXL) and the three task categories, GPT-5.4 AtelierJudge; direct natural-language prompting). Source: arxiv.org. Saturation forecast: Around 2029. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.2 | 65.7 |
| 2 | Gemini 3 Pro (Preview) | 64.7 |
| 3 | Claude Sonnet 4.5 | 64.1 |
| 4 | GPT-4.1 | 63 |
| 5 | Qwen 3 VL 235B A22B | 61.3 |
| 6 | Gemini 2.0 Flash | 60.9 |
| 7 | GPT-4.1 Nano | 60.8 |
| 8 | Qwen 3 VL 8B | 57.8 |
Interactive version: theaggregate.ai/benchmark?slug=ateliereval-image-constraint-satisfaction-novice-prompting · How It Works · Data refreshed daily, snapshot 2026-09-26.