Modes and model IDs
| Mode | Model ID | Inputs & key parameters |
|---|---|---|
| Text-to-image | qwen-image3-t2i-uncensored | prompt; 1K/2K, quality pro/standard, 1–6 images per request, ratios 1:1 · 16:9 · 9:16 · 4:3 · 3:4 |
| Image editing | qwen-image3-edit-uncensored | 1–3 input images + instruction; 1K/2K, pro/standard, 1–6 outputs |
Every mode uses the same request shape: model + prompt + media arrays (images / videos / audios / files as public HTTPS URLs) + params. Full field tables: API docs.
Prices (USD, base list)
| Model | Variant | Price |
|---|---|---|
qwen-image3-edit-uncensored | 1K · Pro quality | $0.04612/image |
qwen-image3-edit-uncensored | 2K · Pro quality | $0.08647/image |
qwen-image3-edit-uncensored | 1K · Standard quality | $0.03459/image |
qwen-image3-edit-uncensored | 2K · Standard quality | $0.03459/image |
qwen-image3-t2i-uncensored | 1K · Pro quality | $0.04612/image |
qwen-image3-t2i-uncensored | 2K · Pro quality | $0.08647/image |
qwen-image3-t2i-uncensored | 1K · Standard quality | $0.03459/image |
qwen-image3-t2i-uncensored | 2K · Standard quality | $0.03459/image |
Base list prices from the live price list (/pricing · pricing.json); the effective price for your key is returned by GET /ent/v2/prices. Cost is reserved at submit and settled on completion — failed tasks are refunded in full.
Example request
POST https://api.18models.ai/ent/v2/generate
Authorization: Bearer sk-…
Content-Type: application/json
Idempotency-Key: <stable-request-id>
{
"model": "qwen-image3-edit-uncensored",
"prompt": "Same woman, same pose; change the outfit to black lace lingerie, keep the lighting",
"images": ["https://cdn.example.com/base.jpg"],
"params": { "resolution": "1k", "quality": "standard", "count": 2 }
}
The response is HTTP 202 with a task_id; poll GET https://api.18models.ai/ent/v2/tasks/{task_id} (back off 2 s → 5 s) until status is success or failed, then read result_urls. Ready-made client code: docs → Build with AI or the Agent Skill.
What is allowed
NSFW output is permitted under the 18models Content Policy: fictional adult content, including explicit sexual content, is allowed; content involving minors (in any style) or real, identifiable people is prohibited absolutely and screened on both input and output. Upstream providers apply their own checks in addition to ours — a task rejected upstream fails and is refunded automatically.
- Editing a photo of a real, identifiable person into sexual content is prohibited (Content Policy §2) and screened on input — use generated characters as the base image.
- Editing adds $0.00346 per input image on top of the per-output price.
FAQ
Does the Qwen-Image API on 18models allow NSFW images?
Yes, for fictional adults, under the 18models Content Policy. Minors and real people's likenesses are prohibited and screened on both input and output.
How much does Qwen-Image 3.0 cost per image?
Standard quality $0.035 per image (1K or 2K), Pro 1K $0.046, Pro 2K $0.086. Editing adds $0.00346 per input image (1–3).
What can Qwen-Image 3.0 editing do?
Instruction-based edits with 1–3 reference images: change clothing, pose, background or style while keeping the subject; batch up to 6 variations per request.
Related
- Z-Image Turbo (cheaper single images)
- Wan 2.7 image (inpainting, grids, 4K)
- API for adult games
- All models & prices
- Pricing
- API docs
- Content Policy