Benchmark RadarAI BENCHMARK PROFILE
WeGenBench
WeGenBench evaluates text-to-image generation across 4,000 bilingual prompts, with multi-dimensional tags and novel VLM-based metrics. It assesses generation quality on scene classification and specific sub-categories.
- Released
- 2026-06-18
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Existing text-to-image benchmarks lack fine-grained diagnostics. WeGenBench provides a structured resource to pinpoint model deficiencies across dimensions, guiding optimization.
Motivation
Recent text-to-image generation models have demonstrated remarkable capabilities in synthesizing highly realistic images from text inputs alone.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.