Benchmark RadarAI BENCHMARK PROFILE
AtelierEval
AtelierEval evaluates prompting proficiency of humans and MLLMs for text-to-image systems via 360 tasks, using a skill-based agentic evaluator (AtelierJudge) that scores prompt-image pairs subjectively and objectively.
- Released
- 2026-05-21
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Introduces a new evaluation angle for T2I pipelines, measuring upstream prompting ability which is currently unassessed.
Motivation
Text-to-image (T2I) systems increasingly rely on upstream prompters, either humans or multimodal large language models (MLLMs), to translate user intent into detailed prompts.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.