Generate with Sogni's current mature-capable video families. Seedance 2.5 Uncensored brings every Seedance 2.5 capability — 4–30 second clips up to 1080p, reference-driven Create, Edit, and Extend — on a more permissive content policy for mature themes, after a one-time likeness and consent agreement, from 30.8 Spark ($0.154) per output second at 480p. Seedance 2.0 Mini Uncensored is the lower-cost Seedance option — 4–15 second clips at 480p or 720p under the same agreement, from 16.8 Spark ($0.084) per output second at 480p. Wan 3 Uncensored — the Wan 3 Enhanced Team’s build of Alibaba’s Wan 3 — produces 2–30 second clips up to 1080p with native audio on a content policy that accepts what standard Wan 3 declines. It is pay as you go on every render, from 16 Spark ($0.08) per output second at 480p. MiniMax H3 is the broad, Seedance 2.0-comparable choice; its 4-step FastH3 Turbo path (FastVideo FastH3 4-step Preview v1) offers up to 6× faster text, image, and first-and-last-frame iteration, and the VBVR Video Reasoning and Mystic X v5 LoRAs add optional prompt control. LTX-2.3 10Eros v1.4 specializes in opt-in image-to-video. All remain subject to platform terms and model requirements.
Seedance 2.5 Uncensored, Seedance 2.0 Mini Uncensored, and Wan 3 Uncensored are pay as you go on every render and are not covered by Unlimited fair use. Unlimited subscribers save 5% and Unlimited Pro subscribers 10% on pay-as-you-go rates, on top of any active promotion.
On this page, uncensored means Sogni supports a wider range of lawful, mature-theme creative work. It does not mean there are no safeguards, licenses, or rules. Sogni's Terms of Use and each model's own requirements still apply. The five model families below are tagged as uncensored in Sogni's current catalog, and all five are revealed by the same explicit 18+ opt-in: switch off the Sensitive Content Filter in your account, then pick the model. They differ in workflow, clip length, and how they are billed.
Seedance 2.5 Uncensored (model ID seedance-2-5-uncensored) is Seedance 2.5 with every capability and limit unchanged on a more permissive content policy: 4–30 second clips up to 1080p, up to 30 image, 10 video, and 10 audio references, and Edit and Extend for source video. Standard Seedance 2.5 declines mature prompts. It is Premium Spark only, from 30.8 Spark ($0.154) per second at 480p, and needs a one-time likeness and consent agreement before the first render.
Seedance 2.0 Mini Uncensored (model ID seedance-2-0-mini-uncensored) is Seedance 2.0 Mini with every capability and limit unchanged on a more permissive content policy: 4–15 second clips at 480p or 720p and up to 9 image, 3 video, and 3 audio references. It is Premium Spark only, from 16.8 Spark ($0.084) per second at 480p, and shares Seedance 2.5 Uncensored’s one-time likeness and consent agreement.
Wan 3 Uncensored is the Wan 3 Enhanced Team’s build of Alibaba’s Wan 3 on a more permissive content policy than the standard build (model ID wan3-uncensored). It is a mature-theme video model and, with Seedance 2.5 Uncensored, one of two that reach 30 seconds at 1080p.
Wan 3 Uncensored is pay as you go on every render — 16 Spark ($0.08) per output second at 480p, 32 Spark ($0.16) at 720p, and 64 Spark ($0.32) at 1080p. With a video reference, the reference seconds bill at those same rates, added to the output seconds. It is a vendor-hosted model, so no part of it falls under Unlimited fair use, but subscribers save 5% (Unlimited) or 10% (Unlimited Pro) on those rates, on top of any active promotion.
For creators searching for an “uncensored Seedance 2.0” alternative, H3 is a strong Sogni match in workflow breadth. FastH3 Turbo adds up to 6× faster iteration. H3 is a separate MiniMax model, not a Seedance derivative.
LTX-2.3 10Eros v1.4 animates one starting image with a mature-audience fine-tune available through the Sogni API and Creative Agent Skill.
Generate dialogue, ambience, effects, and music natively instead of assembling a silent clip and soundtrack in separate tools.
At app.sogni.ai, open your username menu in the top-right corner and switch off Sensitive Content Filter under Content preferences. All generation remains subject to Sogni’s Terms of Use.
H3 Base weights are published under the MiniMax H3 Community License; 10Eros v1.4 is a TenStrip fine-tune with its own source terms. Neither should be described as rule-free.
Create in Sogni Web where supported, through the API, or with Creative Agent—without installing a local video stack.
Seedance 2.5 on a more permissive content policy: 4–30 second video up to 1080p with opt-in mature themes.
Seedance 2.0 Mini on a more permissive content policy: low-cost 4–15 second video at 480p or 720p with opt-in mature themes.
Wan 3 Uncensored — the mature-theme build of Wan 3 — creates 2–30 second video with native audio from text, frame anchors, and multimodal references.
Create polished 5- to 15-second videos with dialogue, music, and sound in one generation. Choose FastH3 Turbo, up to 6× faster than Standard, for everyday creation, or Standard H3 for maximum detail. New two-stage FastH3 output renders 2K at the Standard 768p price.
A mature-theme opt-in LTX-2.3 image-to-video fine-tune for mature-theme creativity, available through the Sogni API and Creative Agent Skill.
MiniMax H3 and H3 Turbo now support two additional community LoRAs. Turn off the Sensitive Content Filter to reveal them in the H3 LoRAs panel.
Improves literal prompt adherence and ordered actions for general or mature-theme work. It supports standard and Turbo H3 text-to-video and image-to-video.
Adds broad mature-theme knowledge while preserving motion, temporal consistency, and fine detail. It supports standard and Turbo text-to-video, image-to-video, and first-and-last-frame video.
At app.sogni.ai, select your username in the top-right corner, switch off Sensitive Content Filter under Content preferences, confirm you are 18 or older, then return to H3 and open LoRAs.
Neither new LoRA is currently offered for H3 reference-to-video. Model licenses, Sogni's Terms of Use, and the opt-in requirement still apply.
| Seedance 2.5 Uncensored | Seedance 2.0 Mini Uncensored | Wan 3 Uncensored | MiniMax H3 family | LTX-2.3 10Eros v1.4 I2V | |
|---|---|---|---|---|---|
| Model relationship | ByteDance’s Seedance 2.5 on a more permissive content policy, with every 2.5 capability and limit unchanged, at a premium rate above standard Seedance 2.5 | ByteDance’s Seedance 2.0 Mini on a more permissive content policy, with every Mini capability and limit unchanged, at a premium rate above standard Seedance 2.0 Mini | The Wan 3 Enhanced Team’s build of Alibaba’s Wan 3 on a more permissive content policy than the standard build | Separate MiniMax model with short-form multimodal scope comparable to Seedance 2.0 | TenStrip mature-audience fine-tune of LTX-2.3 22B |
| Starting point | Text, first frame, first and last frames, image plus audio, or up to 30 images / 10 videos / 10 audio clips as references; Edit and Extend take a source video | Text, first frame, image plus audio, or up to 9 images / 3 videos / 3 audio clips as references; audio needs an image or video reference | Text, first frame, first and last frames, or up to 10 images / 5 videos / 5 audio clips as loose references | Text, one opening frame, first and last frames, or labelled image/video/audio references | One starting image |
| Speed / quality | Vendor-hosted single quality path; no local step or guidance controls | Vendor-hosted single quality path; no local step or guidance controls | Vendor-hosted single quality path; no local step or guidance controls | Standard 20-step quality path, or 4-step FastH3 Turbo (FastVideo FastH3 4-step Preview v1) for up to 6× faster iteration with a possible detail/audio-polish trade-off | Fixed 9-step distilled workflow |
| Clip length on Sogni | 4–30 seconds; Edit follows the source video | 4–15 seconds | 2–30 seconds, explicit output duration | Fixed steps from roughly 5.17–15.08 seconds | 4–20 seconds |
| Output on Sogni | 480p · 720p · 1080p · 24 fps · synchronized native audio | 480p · 720p · 24 fps · synchronized native audio | 480p · 720p · 1080p · 30 fps · synchronized native audio | 768p-class · 24 fps · 32 kHz stereo | 640–3840 px · 1–60 fps (24 default) · native audio |
| Starting price | 30.8 Spark ($0.154) per second at 480p, up to 170.6 Spark ($0.853) at 1080p. Reference-video seconds bill with output seconds, at a lower per-second rate | 16.8 Spark ($0.084) per second at 480p, 37.8 Spark ($0.189) at 720p. Reference-video seconds bill with output seconds, at a lower per-second rate | 16 Spark ($0.08) per billable second at 480p, up to 64 Spark ($0.32) at 1080p. Video references add measured input seconds to output seconds | At 544/768p: Standard 16 Spark ($0.08) · LightX2V Turbo 6 Spark ($0.03) · FastH3 4 Spark ($0.02) per finished second. Reference video adds 16 Spark ($0.08) per input second on every tier | 17.6 Spark ($0.088) for 5 s at 1280 × 720 |
| Plan coverage | Premium Spark on every render; not covered by Unlimited fair use. Subscribers save 5% (Unlimited) or 10% (Unlimited Pro) | Premium Spark on every render; not covered by Unlimited fair use. Subscribers save 5% (Unlimited) or 10% (Unlimited Pro) | Pay as you go on every render; not covered by Unlimited fair use. Subscribers save 5% (Unlimited) or 10% (Unlimited Pro) off pay-as-you-go rates, on top of active promotions | Eligible Unlimited fair-use coverage may apply | Eligible Unlimited fair-use coverage may apply |
| Access | Text, image, reference, Edit, and Extend workflows in Sogni Web, the API, and Creative Agent — filter off, pick Seedance 2.5 Uncensored in a Seedance workflow, and accept the one-time likeness and consent agreement in the Sogni app | Text, image, and reference workflows in Sogni Web, the API, and Creative Agent — filter off, pick Seedance 2.0 Mini Uncensored in a Seedance workflow, and accept the same one-time agreement | Text, image, and reference workflows in Sogni Web, the API, and Creative Agent — filter off, then pick Wan 3 Uncensored under Wan 3 version | All seven standard and Turbo modes through API and Creative Agent; standard t2v/i2v/Ref2VA also in Sogni Web | API and Creative Agent only |
| Optional LoRAs | None — it is a vendor-hosted API model | None — it is a vendor-hosted API model | None — it is a vendor-hosted API model, not a Supernet checkpoint | VBVR Video Reasoning for text/image prompt adherence; Mystic X v5 for text, image, and first/last-frame mature-theme work | No LoRAs currently listed |
| Best for | Mature themes in reference-driven scenes, and editing and extending source video | Mature themes in lower-cost 480p or 720p Seedance takes | Mature-theme clips of up to 30 seconds at 1080p and 30 fps, with dialogue and sound in one pass | General-purpose scenes with broad multimodal control; Turbo for fast iteration, standard H3 for quality-first output | Specialized opt-in mature-theme image animation |
| Model ID | seedance-2-5-uncensored | seedance-2-0-mini-uncensored | wan3-uncensored | Seven workflow-specific standard and Turbo IDs; see the model page | ltx23-22b-10eros-v1.4-fp8mixed_i2v |
These specifications and prices describe Sogni's current deployments, not every upstream Wan, Seedance, H3, or LTX implementation. See each model page for exact duration steps, dimensions, and API examples.
At app.sogni.ai, open your username menu in the top-right corner and switch off Sensitive Content Filter under Content preferences. Confirm you are 18 or older. This reveals every mature-theme model below, including Seedance 2.5 Uncensored and Seedance 2.0 Mini Uncensored in the Seedance workspace and the Wan 3 Uncensored option in the Wan 3 workspace.
Use Seedance 2.5 Uncensored for reference-driven scenes or to edit and extend a source video, and accept its one-time likeness and consent agreement the first time. Use Seedance 2.0 Mini Uncensored for lower-cost Seedance takes at 480p or 720p; the same agreement covers it. Use Wan 3 Uncensored for clips of up to 30 seconds at 1080p and 30 fps — open a Wan 3 workflow and set Wan 3 version to Wan 3 Uncensored. Use standard H3 when quality comes first, FastH3 Turbo for up to 6× faster text/frame iteration, optionally add VBVR Video Reasoning or Mystic X v5, or use LTX-2.3 10Eros for specialized mature-theme animation from a starting image.
Describe the subject, motion, camera, visual treatment, dialogue, ambience, effects, and music. Keep the request realistic for the clip length.
Run Seedance 2.5 Uncensored, Seedance 2.0 Mini Uncensored, and Wan 3 Uncensored from Sogni Web, the API, or Creative Agent. Run every H3 mode through the API or Creative Agent, with standard text, image, and reference modes also in Sogni Web. Run 10Eros through the API or Creative Agent.
Strong video prompts describe what changes after the first frame instead of repeating only what the image already shows. Give the model one coherent shot, name a deliberate camera move, and specify the audio you actually want. Longer prompts are useful when every detail has a job; extra adjectives without direction are not.
Give each upload a role with @Image1, @Video1, or @Audio1; on Seedance 2.5 Uncensored you can also start with “Edit @Video1” and name only what should change. For a real person, attach their photo as a reference; starting from it as the first frame can drift away from the face on big head turns. For mature scenes, attach detailed reference images or videos that show the anatomy and actions you want. Upload only likenesses that are yours or used with explicit, legal consent, and keep every depicted person 18 or older.
Example: “A single unbroken 12-second take in a rain-lit apartment; slow handheld push-in; the woman turns to camera and says ‘You came back.’ Rain on glass, a low room tone, no music.” State wardrobe, framing, and action directly — Wan 3 Uncensored accepts mature-theme subject matter that standard Wan 3 declines, and every depicted person must be 18 or older.
Example: “A ceramic artist opens a glowing kiln; slow push-in; warm documentary lighting. Fire crackles, tools clink softly, and a restrained string cue begins.” Use Turbo for rapid takes and standard H3 for the quality-first pass.
Label each upload and assign its role: <Picture 1> for identity, <Video 1> for motion, or <Audio 1> for voice and sound.
Example: “Slow orbit around the subject as fabric and hair move in a light breeze; hold facial identity; soft room tone and distant city ambience.”
Put spoken words in quotation marks, identify the speaker and voice, and keep the line short enough for the selected duration. Use time ranges only when the shot has multiple beats.
MiniMax documents broader hosted H3 capabilities, but Sogni's current deployment does not expose MiniMax's hosted editing or 2K regeneration stages.
FastH3 Turbo — FastVideo FastH3 4-step Preview v1 — is the default Turbo engine, with LightX2V Turbo one switch away. Turbo paths may trade some fine image detail and audio polish for speed, so compare a standard H3 pass when finish quality matters.
The Sogni deployment requires a starting image, uses a fixed 9-step workflow, and is not currently available in Sogni Web or the native apps.
MiniMax publishes H3 Base weights under its Community License, which includes use and safety conditions. 10Eros has separate creator and base-model terms.
They run on ByteDance’s BytePlus ModelArk rather than the Sogni Supernet, so they are pay as you go on every render — Seedance 2.5 Uncensored from 30.8 Spark ($0.154) and Seedance 2.0 Mini Uncensored from 16.8 Spark ($0.084) per second at 480p — and are not covered by Unlimited fair use. Every account accepts one shared one-time likeness and consent agreement before its first render with either model, and the agreement cannot be accepted with an API key.
It runs on the Wan 3 Enhanced Team’s hosted infrastructure rather than the Sogni Supernet, so it is pay as you go on every render, is not covered by Unlimited fair use, and is bound by the Wan 3 Enhanced Team’s model terms as well as Sogni’s. Subscribers do keep a standing 5% (Unlimited) or 10% (Unlimited Pro) discount on pay-as-you-go rates. “Uncensored” describes a more permissive content policy than standard Wan 3 — not the absence of one.
Verify upstream details in the official MiniMax H3 model card and community license, plus the LightX2V H3 Turbo model card and official Turbo implementation. LTX-2.3 10Eros v1.4 is attributed to TenStrip; see the weights and model notes. Seedance 2.5 Uncensored and Seedance 2.0 Mini Uncensored are ByteDance Seed’s Seedance 2.5 and Seedance 2.0 Mini. Wan 3 Uncensored is the Wan 3 Enhanced Team’s build of Alibaba’s Wan 3; see Alibaba Cloud Model Studio for the base model and its commercial terms. For Sogni-specific IDs, settings, and current pricing, use the Seedance 2.5 Uncensored, Seedance 2.0 Mini Uncensored, Wan 3 Uncensored, MiniMax H3, and LTX-2.3 10Eros model pages.
Need a starting frame for image-to-video? Explore Sogni’s uncensored AI image generator and its 56 current uncensored image models.
Sogni offers 18+ mature-theme workflows marketed as uncensored. Five model families are tagged for that use in Sogni's catalog: Seedance 2.5 Uncensored (model ID seedance-2-5-uncensored), Seedance 2.0 Mini Uncensored (model ID seedance-2-0-mini-uncensored), Wan 3 Uncensored (model ID wan3-uncensored), MiniMax H3 including its three Turbo FL2VA workflows, and LTX-2.3 10Eros v1.4 I2V. H3 also supports the opt-in VBVR Video Reasoning and Mystic X v5 LoRAs. 'Uncensored' does not mean no safeguards: Sogni's Terms of Use and each model's license, provider terms, and safety requirements still apply.
Seedance 2.5 Uncensored is ByteDance's Seedance 2.5 with every capability and limit unchanged, on a more permissive content policy for mature themes, at a premium rate above standard Seedance 2.5; its Sogni model ID is seedance-2-5-uncensored. It generates 4–30 second clips at 24 fps up to 1080p with native audio from text, frames, or up to 30 image, 10 video, and 10 audio references, and can edit or extend a source video. Standard Seedance 2.5 declines mature prompts. Turn off the Sensitive Content Filter, choose Seedance 2.5 Uncensored in a Seedance workflow, and accept the one-time likeness and consent agreement, which covers all Seedance video generation models: you confirm that every likeness you upload is your own or used with explicit, legal consent. API keys cannot accept the agreement, so accept it once in the Sogni app before calling seedance-2-5-uncensored through the API.
Seedance 2.0 Mini Uncensored is ByteDance's Seedance 2.0 Mini with every capability and limit unchanged, on a more permissive content policy, at a premium rate above standard Seedance 2.0 Mini; its Sogni model ID is seedance-2-0-mini-uncensored. It is the lower-cost uncensored Seedance option: 4–15 second clips at 24 fps in 480p or 720p with native audio from text, images, or up to 9 image, 3 video, and 3 audio references, from 16.8 Spark ($0.084) per second at 480p. It uses the one-time likeness and consent agreement for Seedance video generation models, so accepting it once covers them all. For 1080p, clips longer than 15 seconds, or Edit and Extend, use Seedance 2.5 Uncensored.
Wan 3 Uncensored is the Wan 3 Enhanced Team's build of Alibaba's Wan 3 whose content policy accepts mature-theme subject matter standard Wan 3 declines; its Sogni model ID is wan3-uncensored. It generates 2–30 second clips at 30 fps up to 1080p with synchronized native audio, from text, first/last-frame anchors, or loose image, video, and audio references.
Turn off the Sensitive Content Filter first: at app.sogni.ai, select your username in the top-right corner, switch it off under Content preferences, and confirm you are 18 or older. Wan 3 Uncensored then appears under Wan 3 version in any Wan 3 workflow. Through the API or Creative Agent Skill, use wan3-uncensored.
Seedance 2.5 Uncensored and Wan 3 Uncensored. Both reach 30 seconds at 1080p with native audio: Seedance 2.5 Uncensored produces 4–30 second clips at 24 fps and Wan 3 Uncensored 2–30 second clips at 30 fps, against roughly 5–15 second 768p-class clips on MiniMax H3 and 4–20 seconds on LTX-2.3 10Eros. Both are pay-as-you-go on every render rather than eligible for Unlimited fair use.
No. MiniMax H3 is a separate MiniMax model, not a Seedance derivative or modified Seedance checkpoint. It is relevant to searches for an 'uncensored Seedance 2.0' alternative because the models overlap in short-form text, image, reference, dialogue, and native-sound workflows. Sogni supports opt-in mature-content work with H3, but H3 is not restriction-free. For Seedance itself with fewer content restrictions, use Seedance 2.5 Uncensored or Seedance 2.0 Mini Uncensored.
Choose Seedance 2.5 Uncensored for mature themes in clips of up to 30 seconds at 1080p, and for editing and extending source video. Choose Seedance 2.0 Mini Uncensored for lower-cost 480p or 720p clips of up to 15 seconds. Every standard Seedance model also accepts real people with consent after the same one-time likeness agreement. Choose Wan 3 Uncensored for clips of up to 30 seconds at 1080p and 30 fps. For mature scenes, Wan 3 Uncensored needs far less guidance and usually just works, while Seedance needs detailed reference images or videos that show the anatomy and actions you want; Seedance 2.5 Uncensored is the more powerful model and costs more, but it is harder to steer for mature themes. Choose standard MiniMax H3 for quality-first text-to-video, first-frame image-to-video, first-and-last-frame control, or labelled references. Choose FastH3 Turbo for up to 6× faster text, image, or first-and-last-frame iteration in Sogni Web or the API. Choose LTX-2.3 10Eros v1.4 for specialized mature-theme image-to-video.
VBVR Video Reasoning improves literal prompt adherence for standard and Turbo H3 text-to-video and image-to-video. Mystic X v5 adds broad mature-theme knowledge for those modes plus first-and-last-frame video. Neither is currently offered for reference-to-video. Both remain hidden until the Sensitive Content Filter is turned off.
FastH3 Turbo runs the FastVideo FastH3 4-step Preview v1 VSA DataFree checkpoint, a four-step distillation of MiniMax H3, and renders up to 2× faster than the LightX2V 4-step Turbo and up to 6× faster than Standard 768p H3 on a full 15-second clip; shorter clips see a smaller speed-up. Actual time varies by workflow, input, clip length, resolution, worker hardware, and queue state. Turbo paths may trade some fine visual detail and audio polish for speed, and the familiar LightX2V Turbo stays one switch away.
Yes. Seedance 2.5 Uncensored, Seedance 2.0 Mini Uncensored, and Wan 3 Uncensored generate native audio in the same render. MiniMax H3 generates synchronized 32 kHz stereo dialogue, ambience, music, and effects with the picture. LTX-2.3 10Eros can also generate native dialogue and sound in the same pass.
Seedance 2.5 Uncensored and Seedance 2.0 Mini Uncensored are available in Sogni Web, the API, and the Creative Agent Skill, 2.5 Uncensored including Edit and Extend, once their shared one-time agreement is accepted in the Sogni app. Wan 3 Uncensored is available in the same three places, across text-to-video, first/last-frame, and reference-to-video workflows. All seven current MiniMax H3 modes are available through both the Sogni API and Creative Agent Skill: four standard workflows and three Turbo workflows. Sogni Web currently exposes standard H3 text-to-video, image-to-video, and Ref2VA reference-to-video. LTX-2.3 10Eros is available through the API and Creative Agent Skill only, not Sogni Web or native apps.
Seedance 2.5 Uncensored produces 4–30 second clips at 480p, 720p, or 1080p and a fixed 24 fps; Edit follows the source video. Seedance 2.0 Mini Uncensored produces 4–15 second clips at 480p or 720p and a fixed 24 fps. Wan 3 Uncensored produces 2–30 second clips at 480p, 720p, or 1080p and a fixed 30 fps, with an explicit output duration. On Sogni, MiniMax H3 produces roughly 5–15 second 768p-class clips at 24 fps. LTX-2.3 10Eros v1.4 supports 4–20 seconds, dimensions from 640 to 3840 pixels, and 1–60 fps, with 24 fps as the default. Available dimensions and duration steps depend on the selected workflow.
Seedance 2.5 Uncensored costs 30.8 Spark ($0.154) per second at 480p, 69.3 Spark ($0.347) at 720p, and 170.6 Spark ($0.853) at 1080p, with the same subscriber discount as Seedance 2.5 and no Unlimited fair-use coverage. Seedance 2.0 Mini Uncensored costs 16.8 Spark ($0.084) per second at 480p and 37.8 Spark ($0.189) at 720p, with the same discount and no fair-use coverage. With a reference video, either Seedance model bills the reference seconds and the output seconds at one lower per-second rate, so the job still costs more than the same output without one: a 15-second reference plus 15 seconds of 720p output on Seedance 2.5 Uncensored bills as 30 seconds, about 1,244 Spark ($6.22). Wan 3 Uncensored is pay as you go at 16 Spark ($0.08) per billable second at 480p, 32 Spark ($0.16) at 720p, and 64 Spark ($0.32) at 1080p. Without video references, only output seconds are billable. With video references, measured input plus output seconds are billable at those same rates: 10 input seconds plus 5 output seconds at 720p costs 480 Spark ($2.40) before subscriber discounts. It is not covered by Unlimited fair use, and subscribers save 5% (Unlimited) or 10% (Unlimited Pro) on pay-as-you-go rates on top of active promotions. At 544/768p-class output, Standard MiniMax H3 is 16 Spark ($0.08) per finished second, LightX2V Turbo is 6 Spark ($0.03), and FastH3 is 4 Spark ($0.02); in Reference to Video, each reference-video second costs the Standard rate on every tier. An exact 8-second, 192-frame FastH3 clip costs 32 Spark ($0.16). LTX-2.3 10Eros starts at 17.6 Spark ($0.088) for 5 seconds at 1280 × 720 and scales with duration and pixels. Eligible Unlimited fair-use coverage may apply to H3. See the full model coverage table for Unlimited and Premium Spark.
At app.sogni.ai, select your username in the top-right corner, then switch off Sensitive Content Filter under Content preferences. Confirm that you are 18 or older when prompted. In the SDK, use the equivalent filter setting. Mature-theme generation is strictly opt-in.
No. MiniMax H3 and LTX-2.3 10Eros run on compatible Sogni Supernet workers, Seedance 2.5 Uncensored and Seedance 2.0 Mini Uncensored run on ByteDance’s BytePlus ModelArk, and Wan 3 Uncensored runs on the Wan 3 Enhanced Team’s hosted infrastructure. Use Sogni Web where supported, or generate through the API or Creative Agent Skill without installing ComfyUI or owning a local GPU.
Free AI Image Generator · AI Video Generator · Uncensored AI Image Generator · Midjourney Alternative · ChatGPT Alternative · Civitai Alternative · Leonardo AI Alternative · Tensor.Art Alternative · SeaArt.ai Alternative · OpenArt Alternative · Mage.Space Alternative · Venice AI Alternative · Hugging Face Spaces Alternative · ComfyCloud Alternative · Runway Alternative · Higgsfield Alternative · Model Catalog · Sogni Unlimited
Pick Seedance 2.5 Uncensored for 4–30 second clips up to 1080p, Seedance 2.0 Mini Uncensored for lower-cost 480p and 720p takes, or Wan 3 Uncensored for 2–30 second clips up to 1080p, all with native audio and pay as you go — or start with MiniMax H3 in Sogni Web for Unlimited-eligible mature-theme work.