ImageBench V1 —

192 evaluations across 6 categories

Benchmark V1 verdicts are produced by VLM judges and can contain mistakes. Treat PASS/FAIL labels as machine-assisted assessments, and inspect the images yourself. Learn more about the methodology.

Generation Details

Source-backed model context, size, cost, and request settings for this ImageBench V1 run.

local/ming-image-0.1-design-6b

Local

Ming-Image-0.1-Design is Ant Group inclusionAI's 6B MIT-licensed text-to-image model built for text-rich design work: UI screens, infographics, posters and packaging. It pairs a Lumina2/Z-Image-family DiT with the Ling-mini-2.0 MoE prompt encoder, supports native RGBA output, and targets legible long strings and layouts rather than photorealism.

Maker: Ant Group (inclusionAI)
Family: Ming-Image
Model Size: 6B
high
Cost: local run; no API price
not_applicable
Run Target: gx10/ming-image-0.1-design-6b
Effective Request: response_format: b64_json · size: 1024x1024 · num_inference_steps: 12 · seed: 7 · guidance: unguided (BasicGuider, cfg 1.0)
Sources: inclusionAI/Ming-Image-0.1-Design model card inclusionAI/Ming-Image (GitHub)