od Tencent
Hy Image 3.5 Preview
Tencent flagship image generation & editing — 1K–4K tiers, 13 aspect ratios, up to 20 refs
hy-image-v3.5-previewGenerate images with this model.
/api/v1/images/generationsGET /api/v1/models?type=image{
"model": "hy-image-v3.5-preview",
"prompt": "A serene mountain landscape at golden hour",
"size": "1024x1024",
"n": 1,
"reference_image": "https://example.com/reference.jpg"
}curl -X POST https://kunya.ai/api/v1/images/generations \
-H "Authorization: Bearer kunya_YOUR_API_KEY" \
-H "Content-Type: application/json"
-d '{"model":"hy-image-v3.5-preview","prompt":"A serene mountain landscape at golden hour","size":"1024x1024","n":1,"reference_image":"https://example.com/reference.jpg"}'Black Forest Labs
Balanced FLUX 2 Klein — better quality than 4B, still fast
Dokumentacja APIFAL AI (Qwen)
Legacy FAL Qwen LoRA path — use DashScope Qwen Image Max instead
Dokumentacja APIByteDance
ByteDance Seedream 5.0 Pro — high-fidelity text-to-image and editing with 1K/2K tiers, 10 reference images, auto/ratio/custom sizes
Dokumentacja API