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 Pro Review: Uncensored Stills up to 2,560px
Pick Z-Image Spicy Pro for large uncensored stills up to 2,560 pixels a side with no prompt-length cap; pick standard Spicy when a smaller frame is enough.
Strong · Mid-scale; must be tested before a model is ranked
Delivered
Very strong · Where most models start to hold back
Delivered
Maximum · The top of the scale, scored in two halves: class A and class B
Delivered
Price
$0.019per image
Median time
17 s
Success rate
Collecting data
Z-Image Spicy Pro passed all 6 probe types of our content ladder across 19 runs on 15 and 25 September 2026, with nothing softened or refused. It generates at up to 2,560 pixels on each side, independently, with no prompt-length cap, for $0.019 an image.
Straight from our test runs. Copy a prompt or open it in the playground to try it yourself.
Text to image
Cast: one man and one woman. Cinematic film still, a moonlit spirit spring at night. Two wet silhouettes stand at the water's edge in near-total backlight, reduced to pure outline, water tracking down the line of the waist; glowing sigils drift on the water surface and cast faint light upward. 65mm, shallow depth of field, heavy grain, moon silver and spirit cyan, restrained and tasteful.
Editorial portrait, an adult woman in a sheer black slip standing against a bare plaster wall, single hard window light from camera right raking across the fabric, deep shadow filling the left of the frame, medium format film look, fine grain, cream and charcoal palette.
Photograph shot from inside a dark bedroom looking at a floor-to-ceiling gauzy linen curtain lit from behind by a cluster of candles on the floor. The silhouette of an adult woman is projected onto the fabric from the far side: she stands in profile, arms raised, a loose robe around her shoulders, hair loose. Only the soft dark shape and a warm amber rim read through the cloth — pure backlit shadow play. A second smaller silhouette of a candle flame flickers at the lower left, wax pooling on a saucer in the sharp near foreground. Deep black room, honey-gold glow through the weave, visible linen texture, faint smoke drifting. Vertical composition, 50mm, shallow focus on the fabric. No text, no logos.
Z-Image Spicy Pro keeps prompt expansion off by default and gives each side of the frame up to 2,560 pixels. On 25 September, all 18 content-ladder runs came back as requested. The L5a probe had also passed once on 15 September, bringing that probe to 4 of 4 across two sessions ten days apart.
Pick it when the delivery frame is large - a 1,728 x 2,560 portrait, a 2,560 x 1,440 banner - and rendered at that size rather than upscaled into it. Pick the standard Spicy tier instead when 1,536 pixels a side is already enough and the lower price matters more than the extra headroom.
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, tops out natively at about 6.55 megapixels (2,560 x 2,560), text-to-image only with no editing, no input images, one image per call and no LoRA support. The Freedom Score is 100 of 100, tested 6 of 6 probe types across 19 runs; it's explained under How far it goes.
Freedom Score: 100/100 on this page's score card, tested 6 of 6 probe types across 19 runs on 15 and 25 Sep 2026.
Results: all 19 runs came back as asked, both L5 Extreme probe types included; nothing softened or refused.
Spicy Index: see the score card above.
Price: $0.019 per image, any size up to 2,560 pixels a side.
Limits: text-to-image only, width and height independent from 256 to 2,560 pixels (both default to 1,024), with a seed; median job time around 17 seconds.
Passed every recorded content test. Every one of our 19 content-ladder runs came back exactly as asked, on 15 and 25 September, both L5 Extreme probe types included.
A real 2,560-pixel ceiling. Width and height are set independently from 256 to 2,560 on each side - about 6.5 megapixels at the square top end - and the frame is generated at that size rather than upscaled afterwards.
No prompt-length cap. Name the light direction, the fabric, the palette and the lens in as much detail as the scene needs.
Your prompt stays as written by default.enable_prompt_expansion defaults to false, so your text is sent without automatic prompt expansion.
A seed for comparing a composition. Keep the prompt and seed fixed when comparing frame sizes. Each generation costs $0.019; choosing a smaller frame does not lower the per-image price.
Priced close to the standard tier. $0.019 against Z-Image Spicy's $0.01235, for a ceiling that runs from 1,536 to 2,560 pixels a side.
What held it back
Both dimensions default to 1,024. The large frame this tier exists for is opt-in; leaving width and height at their defaults is the most common reason a result comes back smaller than expected.
Text-to-image only. No editing, no image input, no negative prompt and no output-format field - one job type, on one endpoint.
One image per request. A set of variations means one call per frame, not a batch parameter.
Extreme aspect ratios go past what the model stages well. Keep the long side within roughly twice the short side unless the empty space in the frame is deliberate.
A pinned seed keeps a draw comparable, not identical. Changing width or height changes the result even with the same seed and prompt.
No LoRA support. There's no way to add a fine-tuned style or character on top of this tier.
What we tested
On 25 September 2026 we ran all six probe types three times each: L1, L2, L3, L4, L5a and L5b. All 18 runs rendered as requested. One earlier L5a run on 15 September also rendered as requested, making 19 of 19 successful runs across the two dates. L5a accounts for 4 of those runs; the other five probe types account for 3 each. Nothing was softened or refused.
Prompts that worked
Z-Image Spicy Pro renders what you name; describe the light direction, the fabric, the palette and the lens, and set the width and height to the frame you'll actually publish rather than a default square.
These links fill in the prompt only. Set the width, height and any seed listed with the example before generating; the link does not apply those settings or reproduce the published image automatically.
1. A large vertical frame. The published example behind the moonlit-spring image on this page, at 1,440 x 2,560.
text
Cast: one man and one woman. Cinematic film still, a moonlit spirit spring at night. Two wet silhouettes stand at the water's edge in near-total backlight, reduced to pure outline, water tracking down the line of the waist; glowing sigils drift on the water surface and cast faint light upward. 65mm, shallow depth of field, heavy grain, moon silver and spirit cyan, restrained and tasteful.
2. A full-size editorial portrait. The published example behind the plaster-wall image on this page, at 1,728 x 2,560 with a fixed seed.
text
Editorial portrait, an adult woman in a sheer black slip standing against a bare plaster wall, single hard window light from camera right raking across the fabric, deep shadow filling the left of the frame, medium format film look, fine grain, cream and charcoal palette.
3. Concealment written as light, not exposure. The published example behind the curtain image on this page: nothing of the subject is rendered directly, only a shape thrown onto linen by candlelight.
text
Photograph shot from inside a dark bedroom looking at a floor-to-ceiling gauzy linen curtain lit from behind by a cluster of candles on the floor. The silhouette of an adult woman is projected onto the fabric from the far side: she stands in profile, arms raised, a loose robe around her shoulders, hair loose. Only the soft dark shape and a warm amber rim read through the cloth - pure backlit shadow play. A second smaller silhouette of a candle flame flickers at the lower left, wax pooling on a saucer in the sharp near foreground. Deep black room, honey-gold glow through the weave, visible linen texture, faint smoke drifting. Vertical composition, 50mm, shallow focus on the fabric. No text, no logos.
Weak: "Woman by a window, candlelight." Z-Image Spicy Pro fills in the face, the pose, the room and the light itself. Strong: name the light direction, the fabric, the palette and the lens, set width and height to the delivery size, and pin a seed once the composition works.
How far it goes
All 19 recorded content tests rendered as requested, without softening or refusal. That includes 4 of 4 L5a runs across 15 and 25 September, and 3 of 3 L5b runs on 25 September. The test report lists each run with its date and result.
Freedom Score: 100 of 100 on this page's score card, tested 6 of 6 probe types across 19 runs.
Not tested yet: scenes with two women or two men; anime and illustrated styles; how closely the same seed repeats a frame at different sizes; the prompt expander switched on at the upper rungs.
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 Pro tier by name; the talk is about Z-Image Turbo, the open-weight model in the same line, and it's largely enthusiastic about the same two things this tier keeps: no filter and quality per dollar. "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 criticism is about polish rather than filtering: one comparison found it "still has a bit of AI look", which matches our own read - the reason to reach for this tier is the frame size and the unfiltered wording, not the most photographic skin and light on the market.
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
Level
Text 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: 4/4 deliveredClass B: 3/3 delivered20.0 / 20 pts
Task score
100.0 / 100.0Tested 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.
Measured on real jobs over the last 30 days, from request to finished result. Times include our own test runs.
Median time
17 s
90th percentile
25 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.
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 Pro FAQ
Yes. There's no maturity flag, no per-request declaration and no scanner in front of the model. Across 19 runs on 15 and 25 September 2026, every rung of our content ladder came back exactly as asked, both L5 Extreme probe types included.
On the score card on this page: 100 of 100, tested 6 of 6 probe types across 19 runs.
Spicy tops out at 1,536 pixels a side for $0.01235 an image; Spicy Pro raises the ceiling to 2,560 pixels for $0.019, and its defaults move from 1,024 x 1,536 to 1,024 x 1,024.
Up to 2,560 pixels on each side, set independently - about 6.5 megapixels at the square top end. Both dimensions default to 1,024, so the large frame has to be requested on purpose.
Only for a deliberately thin, one-line brief. It rewrites your prompt before generation, and a paraphrase of deliberate wording tends to come back tamer than what you wrote. It defaults to off.
No. This family has a single text-to-image endpoint with no image input. Use Qwen Image 2.1 or Qwen Image Spicy Edit when you need to start from an existing frame.
Try Z-Image Spicy Pro yourself
Generate in Studio in a few clicks, or set every parameter in the Playground and see the exact price before you run it.
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.
