Introduction
What this endpoint does
Qwen Image 2.1 Pro Text to Image renders a written brief into one to six images per request with Alibaba's hosted Pro tier of Qwen Image 2.1. Our requests took 14 to 43 seconds end to end.
Inputs
prompt carries the whole brief, up to 5,000 characters. num_outputs asks for one to six images. aspect_ratio picks one of fifteen shapes and resolution the 1k, 1.5k or 2k tier, all at the same price per image. enable_prompt_expansion (off by default) lets Alibaba rewrite the prompt first, with prompt_optimization_mode and enable_thinking shaping that rewrite. seed steadies a series.
Content
No platform filter. With expansion off, every content test we ran came back as written.
| Field | Value |
|---|---|
| Input | Text |
| Output | Image |
| Status on SpicyAPI | Live: text-to-image and image editing |
| Built on | Qwen Image 2.1, as Alibaba's hosted Pro tier |
| Images per request | 1 to 6, billed per image returned |
| Price | One flat price per image at the 1k, 1.5k and 2k tiers; the live quote is on this page |
| Aspect ratios | 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9, 9:21, 2:1, 1:2, 3:1, 1:3, 5:4, 4:5 |
| 2K sizes | 1:1 2048 × 2048, 16:9 2688 × 1536, 9:16 1536 × 2688, 4:3 2368 × 1760, 3:2 2496 × 1664, 2:3 1664 × 2496 |
| Editing | 1 to 10 reference images; without an aspect ratio the output follows the source at about 2K |
| Prompt | Up to 5,000 characters; no negative prompt, so write exclusions into the prompt |
| Prompt expansion | Off by default; direct or agent rewrite when on, with an optional thinking pass |
| Measured wait | 14 to 43 seconds for text-to-image, 39 to 48 seconds for edits, across our runs |
| Content policy | Uncensored: no platform filter, no maturity flag |
| Weights | Not published; Alibaba offers Pro as a hosted API |
Key features & highlights
A masthead spelled right the first time
Quote the words and say where they go. "AFTER DARK" across the top and the small cover line beneath it came back letter for letter, in both images of the request.
Two portraits, one candlelit table
A man and a woman photographed separately, named as Image 1 and Image 2, end up at one table on a vineyard terrace, both faces still recognisable and her dress changed as asked.
Pro or Turbo, without rewriting a line
Pro and Turbo read the same fields on SpicyAPI, so a prompt that works on one can be sent to the other unchanged. Draft quickly on Turbo and keep the frames that matter on Pro, or the other way round.
Request parameters
| Parameter | Required | Description |
|---|---|---|
prompt | Yes | The whole brief, up to 5,000 characters. Write subject, wardrobe, pose, light and lens in plain sentences. at least 1 characters · at most 5000 characters |
aspect_ratio | No | Fifteen shapes, from 3:1 and 21:9 strips to 1:3 and 9:21 columns. The frame is composed at that shape, not cropped from a square. |
resolution | No | 1k, 1.5k or 2k. Every tier costs the same per image, so 2k is the default. |
num_outputs | No | One to six images from the same prompt in one request. Each image returned is billed, so six images cost six times one. |
enable_prompt_expansion | No | Off by default, so your wording is drawn as written. On, Alibaba rewrites the prompt into a fuller scene first, which helps short everyday prompts but can tone bolder prompts down. Default: |
enable_thinking | No | Spends extra reasoning on the rewrite for a better image, at the cost of a longer wait. Only works while prompt expansion is on. Default: |
prompt_optimization_mode | No | How expansion rewrites the brief when it is on: direct is a single pass that suits most prompts, agent a multi-step rewrite with finer detail. Ignored while expansion is off. |
seed | No | Steadies a draw so you can change one thing at a time and compare fairly. It does not guarantee a pixel-identical image. |
Pricing
$0.036 / image
Use cases
Uncensored image work
SpicyAPI puts no content filter in front of Pro. With prompt expansion off, our tests came back as written at every intensity level we ran, from Mild up to Maximum, with female and male subjects alike.
Covers and posters with exact wording
Alibaba lists improved typography for Pro. Quote the masthead and the cover line, say where each sits, and describe the photograph: our magazine cover came back with both lines spelled exactly, in both images of the request.
One character across a set
Generate a character, then send the frame back as a reference and change the outfit, the room or the light in separate passes. Each edit starts from your approved frame instead of a fresh roll, so the face carries across the set.
Couples from two portraits
Send two portraits as Image 1 and Image 2 and describe the scene they share. In our test a man and a woman photographed separately ended up at one candlelit table on a vineyard terrace, both still recognisable.
Moody portraits and lighting
Alibaba points to portrait lighting and fine detail as Pro's strengths. Name one light source and where it falls, a storm lamp, a high window, a sconce, and the frame is built around it.
Many drafts before one keeper
Up to six images a call at one flat price make a cheap spread of options. Explore wardrobe, pose and light in one request, keep the frame that works, and spend your edits on that one.
Prompt guide & pro tips
Leave prompt expansion off when wording matters
Expansion hands your prompt to Alibaba's rewriter before the image is drawn, and that rewrite can tone a bold brief down. It is off by default here, and every content test we ran with it off came back as written. In our one Pro run with it on, a prompt the official app would block still came through, but keep it off whenever the exact wording matters.
Turn it on for short everyday prompts
A six-word brief gets a fuller scene when expansion is on. direct rewrites in one pass and suits most prompts; agent rewrites in several steps for a finer result. Thinking mode spends extra reasoning on that rewrite and adds time; it does nothing while expansion is off.
Use 2K, it costs the same
Every tier is the same price per image, so 2K is the default and the sensible choice. Drop to 1k only when you want a smaller file or a slightly quicker return.
Ask for several images at once
Set
num_outputsfrom 1 to 6 to get variations of one prompt in one call. Each image that comes back is billed, so six images cost six times one. It is the cheapest way to find a pose or a composition you like before you start editing.Edit with the change first and the keep list second
Write the one thing that changes, then everything that must stay: "Change her velvet dress into an ivory evening gown; keep her face, hair and expression." Name references by number. In our two-portrait test, "the man from Image 1 and the woman from Image 2" at a candlelit table kept both faces recognisable in one frame.
Quote the words that must appear
Put exact wording in quotation marks and say where it goes. Our magazine cover, "AFTER DARK" across the top and "THE SILK ISSUE · NO. 07" in small letters below, came back letter for letter. Keep each line short and check it at full size before you publish.
Write the prompt in a fixed order
Subject, wardrobe, pose, light, lens and palette, in plain sentences. Name the light source and its direction (one oil lamp on the left, a rotating beam behind her) rather than a mood word. With expansion off, Pro draws what you write and does not fill gaps for you.
Introduction
Qwen Image 2.1 Pro Text to Image turns a written brief into images. Pro is the hosted tier of Alibaba's Qwen Image 2.1, which Alibaba announced as officially live alongside the Turbo API on 9 October 2026 and describes as balancing generation quality, inference efficiency and cost, with improved typography, portrait lighting and fine detail. Alibaba offers it as an API and has not published weights for it.
On SpicyAPI it is live. One request returns one to six images, the 1k, 1.5k and 2k tiers cost the same per image, and there is no content filter in front of the model. Every image on this page was generated on Pro with the prompt shown.
Choose a mode by what you have
| Starting material | Mode |
|---|---|
| A written brief and nothing else | Qwen Image 2.1 Pro Text to Image |
| One or more images to change | Qwen Image 2.1 Pro Image Editing |
What we measured
Speed
14 to 31 seconds for a single image and 39 to 43 seconds for two 2K images in one call, end to end across our runs. Prompt expansion and thinking mode add time when they are on.
Content
With prompt expansion off, the default, every content test came back as written, from Mild up to the Maximum level, with a female and a male subject. Not one request was refused. In our one run with expansion on, an unfiltered prompt still came back as asked.
Text in the image
Our magazine cover asked for "AFTER DARK" as the masthead and "THE SILK ISSUE · NO. 07" below it and got both, letter for letter, in both images of the request.
Sizes at 2K
| Ratio | Output size |
|---|---|
| 1:1 | 2048 × 2048 |
| 16:9 | 2688 × 1536 |
| 9:16 | 1536 × 2688 |
| 4:3 | 2368 × 1760 |
| 3:2 | 2496 × 1664 |
| 2:3 | 1664 × 2496 |
Fifteen ratios in all, from 3:1 to 1:3, at 1k, 1.5k and 2k. Output is PNG.
Pro or Turbo
| Pro | Turbo | |
|---|---|---|
| What it is | Alibaba's hosted Pro tier | 8-step checkpoint |
| Weights | Not published | Open, Qwen Research License |
| Images per request | 1 to 6 | 1 to 6 |
| Price by tier | Flat | Flat |
| Price per image | Higher | Lower |
| Content filter | None | None |
The request is the same, so you can send one prompt to both and keep the result you prefer. Alibaba has not published a quality comparison between them.
Prompt guide & Pro tips
Keep expansion off for exact wording
The rewrite can soften a bold brief. Off, Pro draws what you write.
Turn expansion on for short everyday prompts
A six-word idea gets a fuller scene. direct suits most prompts, agent rewrites in more detail, and thinking mode spends extra time on a better rewrite.
Write in a fixed order
Subject, wardrobe, pose, light, lens and palette, in plain sentences. Name the light source and its direction instead of a mood word.
Quote the words that must appear
Put exact wording in quotation marks, keep each line short and say where it sits.
Ask for six, keep one
Variations cost the same each, so a six-image request is the quickest way to find the pose or composition worth editing.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "alibaba/qwen-image-2.1-pro/text-to-image",
"input": {
"prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
"resolution": "2k",
"num_outputs": 1,
"aspect_ratio": "1:1",
"prompt_optimization_mode": "direct"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.
