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.

Z-Image Spicy Review: Uncensored Stills for $0.012 Each

Pick Z-Image Spicy for cheap, fast uncensored stills at an exact size with a seed for sets; skip it for editing, frames over 1,536 px or flawless anatomy.

Spicy Index

73.7/ 100

Freedom Score

98.8/ 100
  • 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
$0.0124 per image
Median time
20 s
Success rate
Collecting data

Our content ladder runs from L1 Suggestive (lingerie, a slipping strap) through L3 Strong and L4 Intense (an intimate scene between two adults) to L5 Extreme, tested here in two probe types. Z-Image Spicy passed all 5 rungs of our content ladder on 25 Sep 2026, one run each, from L1 Suggestive to L5 Extreme, with no refusals and no softened results. It costs $0.01235 per image at any size up to 1,536 pixels.

The verdict

On 15 September we gave Z-Image Spicy our L5 Extreme prompt. Nothing came back covered or toned down, but some of the details we asked for were missing. On 25 September a new L5 run drew every one of them. That is the model in one paragraph: uncensored stills at $0.01235 an image, at the exact pixel size you name, back in seconds, with a seed for building a consistent set, and on the hardest scenes it can take more than one try to land every detail.

Pick it for volume: feed-sized stills, portrait sets, character sheets. Skip it for frames larger than 1,536 pixels, for editing a photo you already have, or when anatomy has to be right on the first try: in our L4 Intense frame, two figures' limbs merged.

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; the full breakdown is in the test report. On specs, it offers free width and height, makes images from text only (no editing, no image input), tops out at about 2.4 megapixels, returns one image per request in a fixed format, and has no LoRA support. The Freedom Score is explained under How far it goes.

Try Z-Image Spicy →

Key facts

  • Freedom Score: on this page's score card, recalculated as runs are added; the extra round on 25 Sep 2026 tested all 5 rungs and every one passed.
  • Results: all 6 runs on 25 Sep 2026 came back as asked, both L5 Extreme probe types included; an L5 run on 15 Sep 2026 had missed some details.
  • Spicy Index: see the score card above.
  • Price: $0.01235 per image, or $0.0133 with prompt expansion on.
  • Limits: text-to-image only, width and height from 256 to 1,536 pixels, with a seed; real jobs take a median of about 18 seconds.
  • Sources: every scored run is in the Z-Image Spicy test report, the scoring rules are on our review methodology page, and the full parameter list is on the Z-Image Spicy model page.

What we liked and what held it back

What we liked

  • Nothing gets toned down. Our 15 September runs came back at exactly the level we asked for: nothing covered, cropped or darkened, and the dim red room light from the prompt intact.
  • It's cheap at any size. $0.01235 an image, $1.24 for a hundred, whether the frame is a 256-pixel thumbnail or 1,536 pixels tall.
  • You name the pixels. Width and height are set independently, anywhere from 256 to 1,536. Generate at the size you'll publish instead of cropping afterwards.
  • A seed for sets. Pin the seed once a frame works, then change one thing per run: the room, then the wardrobe, then the light.
  • It's fast. Real jobs take a median of about 18 seconds, which is what you want when you're drafting a hundred variations.
  • Your words go in as written. Prompt expansion is off by default. When you only have a one-liner, switch it on for $0.00095 more and it writes the scene for you.
  • It stays at the level you ask for. On our 20 September L1 Suggestive brief, Spicy kept the outfit exactly as written; Spicy Pro went further than the brief on its own.

What held it back

  • L5 Extreme is hit and miss. On 15 September the main props made it into the frame, but some details we asked for didn't; on 25 September a new run drew every detail. Spicy Pro drew the full scene in both rounds.
  • Props come out wrong. In one frame a prop we named turned into a long black cord. Describe a prop's shape and material, and check it before you ship.
  • Anatomy slips. One frame exaggerated a figure's proportions far past anything the prompt specified, and in our L4 Intense frame two figures' limbs merged. State the body type you want and check limbs before you ship.
  • Skin looks retouched. Of the three Z-Image tiers on one L1 Suggestive brief, Spicy had the glossiest, most airbrushed skin. For pores and freckles, ask for them by name.
  • Hard light comes out soft. Asked for stripes of sunlight through blinds across her skin, it gave soft window light.
  • Text in, image out, nothing else. No editing, no image input and no negative prompt, and 1,536 pixels is the ceiling on each side.

What the images look like

Our ladder runs aren't shown here; we report the results only. On 15 September, at 1024 × 1536, two L5 Extreme frames came back with nothing covered and the red room light as written. Neither had every detail of the scene we asked for, a prop came out as a cord in one of them, and one exaggerated a figure's proportions well past the prompt. On 25 September, at 1024 × 1024, we ran one prompt per rung plus both L5 Extreme probe types: all six came back as asked, and in the L4 Intense frame two figures' limbs merged.

The gallery holds published, safe-for-work examples from the Z-Image Spicy model page:

The light it softens. A woman in black lace on rumpled white sheets, blinds in the window behind her. The blinds are in frame, but the hard stripes the prompt asked for never land on her.

What expansion adds. The whole prompt was "Adult woman in a silk robe at a rain-streaked window, candlelight."; the brass candlestick and the satin catching the flame came from the expander.

A lived-in scene. A woman on a fire escape at dusk in a lace-trim slip and an open men's shirt, beer and ashtray beside her. Props and setting do the work.

It stops where you tell it to. A man in an open studded leather jacket, lit hard from one side. The prompt ends with a line that sets the limit, and it obeyed.

Prompts that worked

Z-Image draws what you name and fills the rest with its own defaults. Plain words work: the plain names of wardrobe and props render literally, no euphemisms needed. Name five things in this order: who and what they wear, one pose, the light, the lens or film look, the skin. Then set the width and height to the frame you'll publish.

1. The first frame of a set. Based on the prompt behind the blinds image. Once you like a frame, pin the seed and change one thing per run.

text
Photograph of an adult woman lying on her side across rumpled white linen in a black lace bodysuit, morning sun through venetian blinds casting hard stripes across her skin and the sheets, one arm stretched above her head, eyes closed, calm and warm, medium format film look, soft grain, honey and cream palette.

Run this prompt on Z-Image Spicy →

2. A lived-in scene. The exact prompt behind the fire-escape image.

text
35mm film photograph, warm grain and halation, Portra colour. An adult woman sits out on a Brooklyn fire escape at dusk in a black lace-trim slip and an unbuttoned men's dress shirt, bare feet resting through the metal grating, an ashtray and a longneck beer beside her, head tipped back against the railing with her eyes closed. Brick wall, string lights, blue-hour sky, shallow focus.

Run this prompt on Z-Image Spicy →

3. One line, expansion on. Switch on prompt expansion before you run it.

text
Adult woman in a silk robe at a rain-streaked window, candlelight.

Run this prompt on Z-Image Spicy →

Weak: "Woman on a bed by a window." Z-Image picks the face, the outfit, the room and the light. Strong: prompt 1. Same idea, every choice yours, and a seed to hold it.

How far it goes

Z-Image Spicy has not refused a single request. On 25 September we ran every rung of our content ladder once, L1 Suggestive through L4 Intense, plus both L5 Extreme probe types. All six came back as asked. The L4 Intense frame had a quality problem rather than a content one: two figures' limbs merged. The L5 run is the notable change: on 15 September a similar prompt kept the main props but missed some details of the scene, and this time every detail was there. The page label is Uncensored: we add no content filter on our side. The side-by-side with Z-Image Turbo and Spicy Pro is in Is Z-Image uncensored?

Freedom Score: shown on this page's score card and recalculated as we add runs. The extra round on 25 September covered all 5 rungs, one run each, and every one passed.

None of our Z-Image Spicy requests was refused, and every image was billed at the normal price, softened or not.

Not tested yet: the upper rungs with prompt expansion switched on; scenes with two women or two men; anime and illustrated styles; how closely the same seed repeats a frame.

Adults only, in every image. No real, identifiable person should be depicted without their consent.

What the community says

Nobody outside writes about the Spicy tier by name; the talk is about Z-Image Turbo, the open-weight model in the same line. People love it because it's open and unfiltered: "It helps that Z-Image isn't censored", wrote one commenter at launch, and another called it "the first successor to Stable Diffusion 1.5 that delivers better quality". The maker of the GenAI Showdown site said that for a 6B model "it punches far above its weight class", and a daily user said "it has replaced my stock photo subscriptions".

The criticism is about precision and polish: "Not always perfect with text or with following exactly the prompt", and one comparison found it "still has a bit of AI look". That matches our runs: the reason to use it is price, speed and no filter, not the most photographic frame. For that, pay more for Seedream 5.0 Pro.

Alternatives

  • Every L5 detail, every time, or frames up to 2,560 pixels? Z-Image Spicy Pro, $0.019 at any size. It drew every detail in both of our L5 Extreme rounds, and it goes further than an L1 Suggestive brief asks.
  • Cheaper, no seed needed? Z-Image (Turbo), $0.01, five preset shapes. It passed our L3 Strong rung on 11 September.
  • L4 Intense and photo editing? Qwen Image 2.1 passed L4 Intense on 21 and 22 September, and it edits photos, from $0.024. Pony XL did too, for $0.015.
  • Photographic light and skin? Seedream 5.0 Pro, $0.036 at 1K. It put the blind stripes on her skin where Z-Image didn't. Our Seedream 5.0 Pro review.

How far Z-Image Spicy 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
LevelText to image
MildThe lightest prompts in the test
3/3 delivered20.0 / 20 pts
ModerateWhere filtered models start to refuse
3/3 delivered20.0 / 20 pts
StrongMid-scale; must be tested before a model is ranked
3/3 delivered20.0 / 20 pts
Very strongWhere most models start to hold back
3/3 delivered20.0 / 20 pts
MaximumThe top of the scale, scored in two halves: class A and class B
Class A: 3/4 deliveredClass B: 3/3 delivered18.8 / 20 pts
Task score98.8 / 98.8Tested Sep 25, 2026

Tested Sep 15, 2026 to Sep 25, 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

How Z-Image Spicy 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
Not yet tested

No exact match on a tracked public leaderboard yet. How it works

Blind output test 35% of quality
65.3
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
50.0
Freedom Score 30% weight
98.8
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 output2.36 MP
30.0 / 30
TasksText to image
10.0 / 20
Input images per requestNone
0.0 / 20
Shape controlFree width × height
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: 1009 (95% interval 898–1124); votes are still coming in.

What Z-Image Spicy 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
FeatureText to image
Seed controlYes
Negative promptNo
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
20 s
90th percentile
46 s
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.

What people say about Z-Image Spicy

4 positive · 1 mixed · 1 negative

  • It helps that Z-Image isn't censored

    hacker-news· Dec 6, 2025· PositiveSource

  • the first successor to Stable Diffusion 1.5 that delivers better quality

    hacker-news· Dec 6, 2025· PositiveSource

  • it punches far above its weight class

    hacker-news· Dec 7, 2025· PositiveSource

  • it has replaced my stock photo subscriptions

    hacker-news· Jun 24, 2026· PositiveSource

  • Not always perfect with text or with following exactly the prompt

    hacker-news· Dec 6, 2025· MixedSource

  • still has a bit of AI look

    hacker-news· Feb 10, 2026· NegativeSource

Quotes are verbatim from public posts, linked to the original. They are not ratings and don’t feed any score.

Models to consider instead of Z-Image Spicy

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.

Z-Image Spicy FAQ

Try Z-Image Spicy yourself

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

$0.0124 per image

  • 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