Benchmark RadarAI BENCHMARK PROFILE
LingT2I
Evaluates cross-lingual text-to-image generation across content generation and text rendering tasks for 10 languages and 33K prompts.
- Released
- 2026-08-11
- Readiness
- Runnable
- Primary field
- General AI
Why it matters
Reveals cross-lingual inconsistencies in text-to-image models, providing a basis for studying linguistic inequality and improving multilingual generation.
Motivation
Text-to-image (T2I) generation has achieved remarkable progress in recent years.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.