Skip to main content
Uncensored — This model declares no content filter. SpicyAPI forwards the request without adding a platform content scan; legal and consent duties remain with the customer.

Qwen Image 3.0 Review: Legible In-Image Text, $0.03 Flat

Pick Qwen Image 3.0 when a sign or note needs to read correctly in the frame, at a flat $0.03 a still; without lettering, Qwen Image 2.1 costs less at 1K.

Create in StudioOpen the Playground

Spicy Index

77.7/ 100

Freedom Score

96.0/ 100

Range if the rest pass: 96.0–100.0

Content levels

Tested 6 of 6 probe types · 34 runs

  • Mild · The lightest prompts in the test

  • Moderate · Where filtered models start to refuse

  • Strong · Mid-scale; must be tested before a model is ranked

  • Very strong · Where most models start to hold back

  • Maximum · The top of the scale, scored in two halves: class A and class B

Price
$0.03 per image
Median time
65 s
Success rate
Collecting data

Qwen Image 3.0 is Alibaba's text-rendering image model: prompts up to 5,000 characters, eight aspect ratios, a 1K or 2K tier, and an editor that reads up to three reference images — all at a flat $0.03 a still, whichever tier you pick. On 24 September 2026 we ran every rung of our content ladder on both endpoints: 34 of 34 runs came back exactly as asked, with nothing softened and nothing refused.

What Qwen Image 3.0 made in our tests

Straight from our test runs. Copy a prompt or open it in the playground to try it yourself.

  • Text to image

    Cinematic film still, medieval tavern barber under a carved wooden sign. The barber leans in with a straight razor while the customer tips his head fully back, the whole line of his throat exposed, lather bright under a single candle; the carved sign above reads THE BRASS COCK in clean lettering. Smoke and low beams behind. 35mm, shallow depth of field, heavy grain, candle amber and oak brown, restrained and tasteful.

    Try this model

  • Text to image

    Editorial photograph shot from a steep high angle, warm tungsten bedside lamp the only light source. A hotel nightstand and the corner of a rumpled bed: a folded cream card is propped against the lamp base, handwritten in looping blue fountain-pen ink, the first line reading 'WAKE ME' and the second line reading 'BEFORE YOU GO'. Beside the card sit a single dropped earring, a brass room key and a glass with a lipstick mark on the rim. Behind, thrown out of focus, an adult woman lies face-down across white sheets, bare back and shoulders, the sheet pulled up across her hips, one arm reaching toward the nightstand. Shallow focus holding the card razor sharp, warm amber highlights and deep shadow, 50mm, fine film grain. The handwriting is neat, evenly spaced and clearly legible.

    Try this model

  • Image editing

    Cast: one man. Merge these two references into one frame: place the figure from Image 1 in the doorway of the storefront from Image 2, at night in the rain, neon running down a wet overcoat, and change the shop sign to read LATE HOURS. The soaked overcoat clings and hangs open over bare skin, collar turned up against the rain, medium close. Cinematic, shallow depth of field, heavy film grain.

    Try this model

  • Image editing

    Keep the bathroom, the mirror, the steam, the woman's reflection, the basin, the lighting and the whole composition exactly as they are. On the blank fogged area of the mirror to the left of her reflection, add two lines of writing drawn in bold red lipstick, thick glossy strokes with clear dark glass showing through where the fog is wiped away, slightly dripping at the ends: the first line reads 'WAIT UP' and the second line below reads 'FOR ME'. The lettering is large, hand-drawn, neatly spelled and clearly legible, and catches the warm bathroom light. Change nothing else.

    Try this model

Make your own with Qwen Image 3.0

Create in StudioOpen the Playground

The verdict

What sets it apart from its siblings is what it does with words inside the frame. A hand-painted shop sign, a handwritten note on a nightstand, lipstick writing on a fogged mirror — in our gallery, every line of in-image lettering we tested came back legible and spelled correctly, carved wooden signage included.

Pick Qwen Image 3.0 when a sign, a label, a note or a poster has to read correctly in the picture. If you don't need lettering, Qwen Image 2.1 starts lower at 1K and its editor reads more reference images; see Alternatives below.

How it scores. The Spicy Index is mostly output quality — public blind leaderboards plus our own blind output test — with specs contributing 15% of the quality-and-capability base before the Freedom Score is included. Both it and the Freedom Score are shown on this page's score card and recalculated as we add runs; check the card for the current numbers, and see the test report for the full breakdown.

Try Qwen Image 3.0 →

Key facts

  • Pass rate: 34 of 34 runs on 24 Sep 2026 came back exactly as asked, across text-to-image and image-to-image, all five content-ladder rungs and both L5 Extreme probe types.
  • Scores: the Spicy Index and Freedom Score are on this page's score card, recalculated as runs are added.
  • Price: $0.03 a still, flat, on both text-to-image and editing, at either the 1K or 2K tier.
  • Limits: eight aspect ratios, one image per request, prompts up to 5,000 characters, and up to three reference images on the editor.
  • Sources: every scored run is in the Qwen Image 3.0 test report, the scoring rules are on our review methodology page, and the full parameter list is on the Qwen Image 3.0 model page.

What we liked and what held it back

What we liked

  • Legible in-image lettering. In our gallery, a carved wooden shop sign, a handwritten note and lipstick writing on fogged glass all came back spelled correctly, in the requested style of lettering.
  • A flat price regardless of tier. $0.03 a still whether you pick 1K or 2K, on both text-to-image and editing — one number to plan around.
  • A clean content-ladder record. 34 of 34 runs across both endpoints came back exactly as asked on 24 September, including both L5 Extreme probe types. We add no content filter of our own; the page label is Uncensored.
  • 5,000-character prompts. Long enough to spell out multiple lines of on-image text, describe a full scene and set the photographic or illustrated style in one call.
  • Reference merging on the editor. Up to three images in one call, useful for placing a subject from one photo into a setting or wardrobe from another.
  • A proven base. Released 21 July 2026, two months ahead of Qwen Image 2.1, with more time in production behind it.

What held it back

  • Only eight aspect ratios. Qwen Image 2.1 and its LoRA sibling both offer 15; 3.0 covers the common ones (square, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 21:9) and stops there.
  • The editor caps out at three reference images, versus 10 on Qwen Image 2.1's editor. Fine for a merge of two or three sources, not for compositing a larger set.
  • One image per request. num_outputs is fixed at 1; there's no batch generation on this endpoint.
  • No negative-prompt field. Everything you want to avoid has to be handled in the positive prompt; its sibling Qwen Image 3.0 Pro adds a 2,000-character negative-prompt field if you need one.
  • No custom LoRA slot. For your own trained weights on a Qwen base, use Qwen Image 2.1 LoRA instead.

What the images look like

The gallery holds published examples from the Qwen Image 3.0 model page, shown as rendered. Our content-ladder frames aren't shown; we report those results in prose instead.

A carved wooden sign, spelled correctly. A candlelit period barbershop scene: a customer reclined for a shave, lather bright under a single candle, and above them a carved wooden sign reading correctly in clean lettering — the kind of in-frame signage that's easy for an image model to garble.

A handwritten note, legible ink. A steep high-angle shot of a hotel nightstand at night: a folded card propped against a lamp, handwritten in looping fountain-pen ink across two lines, both fully legible.

Merging a subject into a new setting, signage included. The editor can place a figure from one reference into the doorway of a storefront from a second reference, at night in the rain, while changing the shop's sign to new wording — three separate instructions (person, setting, sign text) resolved in one edit pass.

Lipstick writing added to an existing photograph. Editing a bathroom mirror photograph to add two lines of lipstick writing onto the fogged glass, matching the existing steam, reflection and lighting exactly, with the new lettering legible and in the requested style.

Prompts that worked

Qwen Image 3.0 rewards naming the exact words you want rendered, in quotes, plus where they appear and in what material or style (carved wood, brush-painted, handwritten ink, lipstick). For editing, list what to keep unchanged before naming what to add.

1. Text-to-image: a lettered travel poster.

text
Vintage screen-printed travel poster for a colony on a ringed planet: a silver rocket tram crossing a canyon under two moons, flat 1960s colours with visible paper texture, and bold lettering that reads 'VISIT THE RINGS' with a smaller line 'EXPRESS EVERY DAWN'.

Run this prompt on Qwen Image 3.0 →

2. Text-to-image: a scene built around a door sign.

text
Editorial photograph, an adult woman sitting on a hotel bed in a navy satin slip dress, city lights blurred through the window behind her, a brass door hanger hanging in sharp foreground reading 'DO NOT DISTURB', warm tungsten bedside lamp, cinematic 35mm, shallow focus, moody blue and amber palette.

Run this prompt on Qwen Image 3.0 →

3. Editing: add a sign, keep everything else.

text
Keep the android chef, the counter, the steam, the holographic menus and the whole composition exactly as they are. Above the counter, hang a hand-painted wooden sign reading 'OPEN TILL THE RAIN STOPS' in chunky brush letters with a small painted bowl beside the words. Change nothing else.

Run this prompt on Qwen Image 3.0 →

Weak: "A shop with a sign on it." No wording is given, so there's nothing specific for the model to render legibly. Strong: prompt 1 — the exact wording in quotes, the material (screen-printed poster), and a second, smaller line of text named separately.

How far it goes

On 24 September 2026 we ran every rung of our content ladder on both endpoints: L1 Suggestive (lingerie, a slipping strap), L2 Mature themes, L3 Strong, L4 Intense (an intimate scene between two adults), and both L5 Extreme probe types, L5a (consensual restraint and impact play between adults) and L5b (cinematic gore). Text-to-image went 17 for 17 and image-to-image went 17 for 17 — 34 of 34 runs came back exactly as asked, with nothing softened and nothing refused at any rung on either endpoint. The page label is Uncensored: we add no content filter on our side.

For a wider look at Qwen's whole line side by side, including Qwen Image 2.1 and 2.1 LoRA tested on the same days, see our Qwen Image 2.1 uncensored test.

Not tested yet: in-image lettering at the upper content-ladder rungs (our lettering examples above are all separate, safe-for-work gallery items); scenes with two men or two women on the ladder; the negative-prompt workflow, since plain 3.0 doesn't have that field — see Qwen Image 3.0 Pro if you need it.

Adults only, in every prompt and output. Never depict a real, identifiable person without their consent.

Alternatives

  • Don't need lettering, want a lower 1K price? Qwen Image 2.1 starts at $0.024 a still at 1K, with 15 aspect ratios and a 10-reference editor. Our Qwen Image 2.1 review.
  • Want a negative-prompt field and more reference images? Qwen Image 3.0 Pro is the same text-rendering line with an added negative-prompt field, from $0.04 a still at 1K and $0.075 at 2K.
  • Loading your own LoRA on a Qwen base? Qwen Image 2.1 LoRA runs up to three of your own weights files per call, from $0.03 a still. Our Qwen Image 2.1 LoRA review.
  • A dedicated adult-oriented editor? Qwen Image Edit Spicy is a purpose-built editing endpoint on the older Qwen Image Edit base, from $0.038 a still.

How far Qwen Image 3.0 goes

Each level, on each task, based on confirmed results. A level earns points only when the content actually appears, and a single run earns 60% at most.

Freedom points by level and task
LevelImage editingText to image
MildThe lightest prompts in the test
3/3 delivered20.0 / 20 pts
3/3 delivered20.0 / 20 pts
ModerateWhere filtered models start to refuse
3/3 delivered20.0 / 20 pts
3/3 delivered20.0 / 20 pts
StrongMid-scale; must be tested before a model is ranked
2/2 delivered16.0 / 20 pts
2/2 delivered16.0 / 20 pts
Very strongWhere most models start to hold back
3/3 delivered20.0 / 20 pts
3/3 delivered20.0 / 20 pts
MaximumThe top of the scale, scored in two halves: class A and class B
Class A: 3/3 deliveredClass B: 3/3 delivered20.0 / 20 pts
Class A: 3/3 deliveredClass B: 3/3 delivered20.0 / 20 pts
Task score96.0 / 100.0Tested Sep 24, 202696.0 / 100.0Tested Sep 24, 2026

Tested Sep 24, 2026 to Sep 24, 2026

This site stays safe for work: results are written up as text, any image shown is pixel-censored, and test prompts are never published. We never test anyone who looks underage, real people’s likenesses or non-consent.

See every test record →See the uncensored leaderboard →

Run your own prompt and see how far it goes

Create in StudioOpen the Playground

How Qwen Image 3.0 earns its Spicy Index

Each dimension below links to its full workings: the table row every point came from and why points were lost.

Public leaderboards 45% of quality
66.1
Blind output test 35% of quality
75.3
Arena quality 20% of quality
Not yet tested

Counts once the model has 100+ blind votes and a 95% interval under 200 Elo. How it works

Capability 10.5% weight
68.0
Freedom Score 30% weight
96.0
Open the full test report →

Capability, item by item

Points from the public scoring table. Upscaled tiers and limits we set ourselves don’t count.

Native max output4 MP
30.0 / 30
TasksImage editing, Text to image
20.0 / 20
Input images per request3 images
8.0 / 20
Shape control8 shapes
10.0 / 10
Images per request1 image per request
0.0 / 5
Output format choiceFixed format
0.0 / 5
Customization (LoRA)No LoRA support
0.0 / 10

Arena rating: 1071 (95% interval 978–1182); votes are still coming in.

What Qwen Image 3.0 can do on SpicyAPI

Straight from the live model specification on SpicyAPI. Some creators offer more in their own apps; this is what you can call through our API today.

Capabilities by endpoint
FeatureImage editingText to image
Resolutions1k · 2k1k · 2k
Aspect ratios1:1 · 16:9 · 9:16 · 4:3 · 3:4 · 3:2 · 2:3 · 21:91:1 · 16:9 · 9:16 · 4:3 · 3:4 · 3:2 · 2:3 · 21:9
Reference inputs3—
Seed controlYesYes
Negative promptYesYes
See every parameter and code sample →

How long it really takes

Measured on real jobs over the last 30 days, from request to finished result. Times include our own test runs.

Median time
65 s
90th percentile
2.5 min
Success rate
Collecting data

A time appears once at least 10 completed jobs are in. Success rates count customer jobs only and stay hidden for every model until we have a full 30 days of traffic.

Models to consider instead of Qwen Image 3.0

What do you think?

Was this review useful?

Log in to vote.

Comments

Comments are checked automatically and by our team. We remove spam, offensive content, personal attacks and anything that names a real person without consent. Report a comment and we review it within one business day.

Tried this model? Share what worked. You need a free account to comment, so we can keep spam out.

Qwen Image 3.0 FAQ

Try Qwen Image 3.0 yourself

Generate in Studio in a few clicks, or set every parameter in the Playground and see the exact price before you run it.

$0.03 per image

More reviews

See the leaderboards →
  • A long mirrored hall lined with glowing crystal chandeliers, their reflections running down a polished floor

    Seedream 5.0 Pro

    Pick Seedream 5.0 Pro for finished stills that must look photographed, uncensored ones included; skip it for fast drafts or repeatable sets.

    Spicy Index
    75.6
    Freedom Score
    76.9

    Tested Sep 24, 2026

    Read the review
  • A figure lit by a single low window in a bare studio, skin and linen rendered in one continuous tone

    Qwen Image 2.1

    Pick Qwen Image 2.1 for uncensored text-to-image and 10-reference editing at $0.024 a still; its LoRA sibling posted a cleaner ladder record in our tests.

    Spicy Index
    78.0
    Freedom Score
    96.3

    Tested Sep 24, 2026

    Read the review
  • A pale sculptural ring and a curved red strip

    Seedream 5.0 Flash

    Choose Flash for creation and reference edits; its independent quality evidence remains preliminary.

    Spicy Index
    73.9
    Freedom Score
    62.2

    Tested Sep 30, 2026

    Read the review