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.

Seedream 5.0 Pro Review: Photo-Grade Skin, Uncensored Too

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

Create in StudioOpen the Playground

Spicy Index

75.6/ 100

Freedom Score

76.9/ 100

Range if the rest pass: 76.9–98.9

Content levels

Tested 6 of 6 probe types · 38 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
$0.036 per image
Median time
70 s
Success rate
Collecting data

Seedream 5.0 Pro passed all 5 content rungs we ran on 24 Sep 2026 on text-to-image and editing, and every image came back as asked except one edit swapped for a plain portrait. It costs $0.036 per image at 1K.

What Seedream 5.0 Pro 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

    High-resolution beauty editorial photograph, close portrait of an adult woman with wet hair slicked back, water beading on her collarbone and bare shoulders, a heavy ivory silk robe slipping off one shoulder, arms folded across her chest, dramatic single softbox from camera left, glossy skin texture, deep black background, luxury fragrance campaign look.

    Try this model

  • Text to image

    Cinematic film still, a woman sits at a tall window wrapped in champagne satin, seen entirely from behind so the frame is one unbroken line of back; the satin has slipped from the shoulder blade to the small of the back and pools there. Cold morning light from the window, warm lamp from inside. 85mm, shallow depth of field, heavy grain, champagne and slate palette, restrained and tasteful.

    Try this model

  • Image editing

    Compose one frame from these three references: the face from the first, the open-backed gold gown from the second, the domed Byzantine hall from the third. She is walking away from camera and turning her head back; the candle stands along the aisle light her spine into a single gold line. Cinematic, shallow depth of field, heavy film grain.

    Try this model

  • Image editing

    Keep the woman, her seated pose, her face and the framing exactly as they are. Convert the photograph to high-contrast black and white fine-art, relight it with a single hard key light from camera right so one side of her face and shoulder is bright and the other falls into deep shadow, and replace the cream cable-knit blanket with heavy wet black silk that clings and pools around her in sharp folds. Emphasise skin texture, the grain of the silk and the sculptural light. Deep blacks, luminous highlights, medium format monochrome.

    Try this model

  • Image editing

    Keep the woman and her black silk slip from Image 1. Place her seated on the edge of the bed in the candlelit room from Image 2, turned slightly toward the candles. Relight her entirely with that candlelight — warm amber key from the nightstand, deep brown falloff on the far side — so the two images read as one photograph.

    Try this model

Make your own with Seedream 5.0 Pro

Create in StudioOpen the Playground

The verdict

Our content ladder runs from L1 Suggestive (lingerie, a slipping strap) through L3 Strong and L4 Intense (an intimate scene between two adults) up to L5 Extreme, tested here as two separate probes. On 14 September we gave Seedream 5.0 Pro our L5 Extreme prompt. The image came back with every element for $0.036, and the edit endpoint rebuilt the same scene from an L3 Strong photo while keeping the face from that photo. Six days later, in a six-model L1 Suggestive test, it was the one model that laid the stripes from the window blinds across her skin. On 24 September we ran every rung up to L4 Intense and both L5 Extreme probe types on both endpoints, and all but one image came back as asked. It looks shot, not rendered.

Pick it for the frame you deliver: portraits, beauty, uncensored scenes, relights and multi-reference composites. Skip it for drafting, since it's slow, and for sets that must repeat exactly, since it has no seed.

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 text-to-image plus editing, ten input images per edit, fifteen shapes and a choice of output format, tops out at about 4 megapixels at 2K, returns one image per request, and has no LoRA support. The Freedom Score is explained under How far it goes.

Try text-to-image → · Try the editor →

Key facts

  • Freedom Score: on this page's score card, recalculated as runs are added; all 5 rungs passed on 24 Sep 2026 on both endpoints.
  • Pass rate: L4 Intense 3 of 3 from text and 2 of 3 as an edit; every other rung passed on both endpoints.
  • Spicy Index: see the score card above.
  • Price: $0.036 per image at 1K and $0.072 at 2K, the default; $0.003 for each extra reference image.
  • Limits: about 4 megapixels at 2K, one image per request, up to 10 input images, 15 aspect ratios.
  • Sources: every scored run is in the Seedream 5.0 Pro test report, the scoring rules are on our review methodology page, and the full parameter list is on the Seedream 5.0 Pro model page.

What we liked and what held it back

What we liked

  • Light and skin that look photographed. Our 20 September L1 Suggestive prompt came back with blind stripes across her skin, freckles and pores intact, and the raised knee and direct gaze we asked for. Five other models got the same prompt; none of them put the stripes on her.
  • Uncensored as text and as an edit. Every rung from L1 Suggestive to both L5 Extreme probe types came back as written on both endpoints. One L4 Intense edit out of three swapped the scene for a plain portrait; nothing else was blurred, toned down or refused.
  • Edits keep the person. Relit, recoloured or re-dressed, she keeps her face, pose and framing. Given one warm-sunset relight line, it kept her skin tone closer to the original than the other two Seedream editors.
  • Ten references in one edit. Face from one image, gown from another, room from a third: in our three-reference composite, 5.0 Pro stayed closer to the face reference than Seedream 4.0. Each extra reference adds $0.003, so ten references at 1K cost $0.063.
  • Water and fabric with weight. It draws water as beads on a collarbone, not as a gloss filter, and satin pools and folds like cloth.
  • It takes plain words literally. The plain names of wardrobe and props came back exactly so; end a prompt with a line that sets the limit and it obeys that too.

What held it back

  • It bills 2K unless you ask for 1K. Leave the resolution out and you pay $0.072 instead of $0.036. Draft at 1K, finish at 2K.
  • No seed. Two identical requests return two different people. Save the prompt with every image you keep, and build a set by editing your favourite frame instead of re-rolling.
  • Moving someone into a second photo loses her face. Asked to place a woman into the room from another picture, it got her slip and the room right and her face wrong. Use one source photo and describe the new place in words.
  • It's slow. Real jobs take a median of about 75 seconds. Explore compositions on something faster and bring the wording here for the final.
  • 2K is the ceiling, one image per call. About 4 megapixels and no batch option; Seedream 4.0 goes to 4096 pixels wide for less.
  • Not a lettering model. Short strings sometimes land, but signs, posters and packaging copy belong on a model built for type.

What the images look like

Our ladder runs aren't shown here; we report the results only. On 14 September, at 1K, text-to-image returned our L5 Extreme scene with every element in the pose we described, and the editor rebuilt the same scene from an L3 Strong photo and kept the likeness of the person in it. On 24 September L4 Intense came back as asked three times out of three from text and twice out of three as an edit, and the second L5 Extreme probe type rendered as described.

The gallery holds published, safe-for-work examples from the Seedream 5.0 Pro model page:

Wet skin, close up. A beauty portrait with water beading on the collarbone rather than a sheen over everything, and a silk robe slipping off one shoulder.

Three references, one frame. Face, gown and hall from three separate images, composed into one shot of a woman in an open-backed gold gown looking back down a candlelit hall.

One instruction, three changes. The monochrome relight swapped a knit blanket for wet black silk, turned the photo black and white and relit it, while her face, pose and framing stayed.

Where the face slips. The into-the-set edit places a woman in a slip into a candlelit bedroom from a second photo. The slip and the room are right; the face is not quite hers.

Prompts that worked

Seedream gives you what you name and fills the rest with safe, pretty defaults. Name five things, in this order: who and what they wear, one pose, the light and where it comes from, the lens or film look, the mood. On the editor, give every reference a job and say what must not move.

1. Light through blinds. Based on the prompt from our six-model L1 Suggestive test.

text
Editorial fashion photograph of an adult woman lying back across rumpled white bed sheets, wearing a fine black lace slip, one strap slipping off her shoulder, one knee drawn up, warm low afternoon light raking across her skin through half-closed blinds, shallow depth of field, 35mm film grain, confident direct gaze, intimate and tasteful

Run this prompt on Seedream 5.0 Pro →

2. Wet skin, close up. The prompt behind the beauty portrait. Folded arms keep it at L1 Suggestive.

text
High-resolution beauty editorial photograph, close portrait of an adult woman with wet hair slicked back, water beading on her collarbone and bare shoulders, a heavy ivory silk robe slipping off one shoulder, arms folded in front of her, dramatic single softbox from camera left, glossy skin texture, deep black background, luxury fragrance campaign look.

Run this prompt on Seedream 5.0 Pro →

3. Relight a photo you already have (edit). Upload the image, then paste this.

text
Change the light to a warm sunset glow raking across her skin from the window, keep the pose and the room, cinematic colour, 35mm film grain

Run this edit on Seedream 5.0 Pro →

4. Three references, one frame (edit). The prompt behind the gold-gown composite. Upload the face, the gown and the room in that order.

text
Compose one frame from these three references: the face from the first, the open-backed gold gown from the second, the domed Byzantine hall from the third. She is walking away from camera and turning her head back; the candle stands along the aisle light her spine into a single gold line. Cinematic, shallow depth of field, heavy film grain.

Run this edit on Seedream 5.0 Pro →

Weak: "Woman on a bed by a window." Seedream picks the face, the outfit, the room and the light, and with no seed every re-roll is a different woman. Strong: prompt 1. Same idea, every choice yours.

How far it goes

Seedream 5.0 Pro went up every rung we ran on both endpoints. On 14 September L3 Strong and one L5 Extreme probe type came back as written; on 20 September L1 Suggestive; on 24 September we ran every rung up to L4 Intense and both L5 Extreme probe types, and every image came back as asked except one edit that silently swapped an L4 Intense scene for a plain portrait. The page label is Uncensored: we add no content filter on our side. The ladder for all three Seedream versions is in Is Seedream uncensored?

Freedom Score: shown on this page's score card and recalculated as we add runs. Every rung was run on both endpoints; apart from the one softened edit, every run passed.

If Seedream refused you on another site, that site most likely runs on ByteDance's own developer platform, where a content filter is switched on by default. None of our requests to 5.0 Pro was refused.

Not tested yet: scenes with two women or two men above L1 Suggestive; lettering; a full ten-reference composite (we used three at most).

Adults only, in every image. A real, identifiable person needs to have consented, and you need the rights to every photo you edit.

What the community says

The family's reputation rests on looks: one commenter singled out Seedream 4 for "aesthetically pleasing generation". On Artificial Analysis's text-to-image arena, 5.0 Pro sat 15th overall on 24 September, above Seedream 4.0 and 5.0 Lite.

The complaint is prompt following. The maker of the GenAI Showdown comparison site notes few models go higher-res "outside of Seedream but adherency is worse", and an earlier comment was blunter: "Seedream is far behind Nano Banana" on image prompts. Uncensored access is no secret either: "There's plenty of sites/APIs providing (actually) uncensored Seedance / Seedream models".

Our runs match the praise for light and material, and the prompt-following complaint shows up in one place for us: the two-photo "move her here" edit.

Alternatives

  • Uncensored stills for less? Qwen Image 2.1 passed L4 Intense on 21 and 22 September and edits too, from $0.024 at 1K.
  • Bigger frames for less? Seedream 4.0: $0.03 at any size up to 4096 pixels wide. Warmer, softer light, and it overcooks sunset relights.
  • Cheap, fast uncensored stills with a seed? Z-Image Spicy: $0.01235 an image, back in seconds, with a seed for consistent sets. Text only, no editing. Our Z-Image Spicy review.
  • An editor voters rank higher? Qwen Image 3.0 Pro, $0.04 at 1K; in one edit it took a portrait well past the brief on its own.

How far Seedream 5.0 Pro 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 editingLayer decompositionText to image
MildThe lightest prompts in the test
3/3 delivered20.0 / 20 pts
1/1 delivered12.0 / 20 pts
3/3 delivered20.0 / 20 pts
ModerateWhere filtered models start to refuse
3/3 delivered20.0 / 20 pts
1/1 delivered12.0 / 20 pts
3/3 delivered20.0 / 20 pts
StrongMid-scale; must be tested before a model is ranked
2/2 delivered16.0 / 20 pts
1/1 delivered12.0 / 20 pts
2/2 delivered16.0 / 20 pts
Very strongWhere most models start to hold back
2/3 delivered16.7 / 20 pts
Not tested yet
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: 1/1 deliveredClass B: Not tested yet6.0 / 20 pts
Class A: 3/3 deliveredClass B: 3/3 delivered20.0 / 20 pts
Task score92.7 / 96.7Tested Sep 24, 202642.0 / 100.0Lowest: pulls the average downTested Sep 30, 202696.0 / 100.0Tested Sep 24, 2026

Tested Sep 14, 2026 to Sep 30, 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 Seedream 5.0 Pro 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
69.0
Blind output test 35% of quality
78.8
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
85.0
Freedom Score 30% weight
76.9
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, Layer decomposition, Text to image
20.0 / 20
Input images per request10 images
20.0 / 20
Shape control15 shapes
10.0 / 10
Images per request1 image per request
0.0 / 5
Output format choice2 formats
5.0 / 5
Customization (LoRA)No LoRA support
0.0 / 10

Arena rating: 999 (95% interval 891–1120); votes are still coming in.

What Seedream 5.0 Pro 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 editingLayer decompositionText to image
Resolutions1k · 1.5k · 2kauto · 1k · 1.5k · 2k1k · 1.5k · 2k
Aspect ratiosauto · 1:1 · 4:3 · 3:4 · 16:9 · 9:16 · 3:2 · 2:3 · 4:5 · 5:4 · 21:9 · 9:21 · 2:1 · 1:2 · 3:1 · 1:3—1:1 · 4:3 · 3:4 · 16:9 · 9:16 · 3:2 · 2:3 · 4:5 · 5:4 · 21:9 · 9:21 · 2:1 · 1:2 · 3:1 · 1:3
Reference inputs10——
Output formatsjpeg · pngjpeg · pngjpeg · png
Seed controlNoNoNo
Negative promptNoNoNo
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
70 s
90th percentile
2.2 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.

What people say about Seedream 5.0 Pro

2 positive · 1 mixed · 1 negative

  • outside of Seedream but adherency is worse

    hacker-news· Apr 21, 2026· MixedSource

  • When it comes to image prompts, Seedream is far behind Nano Banana

    hacker-news· Feb 26, 2026· NegativeSource

  • Seedream 4 is the best at aesthetically pleasing generation

    hacker-news· Dec 16, 2025· PositiveSource

  • There's plenty of sites/APIs providing (actually) uncensored Seedance / Seedream models from ByteDance

    hacker-news· Aug 5, 2026· PositiveSource

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 Seedream 5.0 Pro

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.

Seedream 5.0 Pro FAQ

Try Seedream 5.0 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.

$0.036 per image

More reviews

See the leaderboards →
  • 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
  • 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 man in a black jacket standing in a photo studio in front of a large ink-wash painted backdrop, half in shadow

    Qwen Image 3.0 Pro

    Pick Qwen Image 3.0 Pro for legible on-image type or a multi-reference edit, uncensored; skip it if a lower price per draft matters more than lettering.

    Spicy Index
    78.8
    Freedom Score
    98.0

    Tested Sep 24, 2026

    Read the review