Wan 2.7 Uncensored
Alibaba's flagship 27B open-weight video model — uncensored, featuring native audio synchronization and exceptional character consistency.
Overview
What is Wan 2.7 Uncensored
Wan 2.7 Uncensored is an advanced, open-weight video model (text-to-video and image-to-video) developed by Alibaba's Tongyi Lab and released in April 2026. Built on a 27-billion-parameter architecture, it delivers high-fidelity video generation with native audio synchronization, exceptional character consistency, and complete creative freedom without corporate censorship.
Running it privately on Venice
On Venice, you can run Wan 2.7 Uncensored privately with zero prompt retention or surveillance. Unlike centralized platforms that monitor and restrict your creative output, Venice provides a permissionless, uncensored environment where your prompts are anonymized. You maintain complete sovereignty over your video generation workflows.
Assessment
Strengths and limitations
- Native audio synchronization that aligns sound effects and musical cues with on-screen motion in a single pass.
- Uncensored and permissionless, allowing creators to generate videos without arbitrary corporate guardrails.
- Open weights under Apache 2.0 license, ensuring transparency and self-hosting capabilities.
- Strong subject and character consistency across generated clips.
- Supports up to 15-second clips at 1080p resolution with multiple aspect ratios.
- Requires significant computational resources if run locally due to its 27B parameter size.
- Text-to-video generation can occasionally require prompt iteration for highly complex physical interactions.
- Served via third-party providers on Venice, meaning requests are anonymized but processed off-site.
Capabilities
What it supports
- Text to video
- Native audio generation
Specifications
Datasheet
- Maker
- Alibaba Tongyi Lab
- Released
- April 3, 2026
- Architecture
- 27-billion parameter video diffusion model
- Modality
- Text-to-video with native audio
- Open weights
- Yes — Apache 2.0
- Resolutions
- 1080p, 720p
- Clip lengths
- 5s, 10s, 15s
- Mode
- text-to-video
- Aspect ratios
- 16:9, 9:16, 1:1
- Audio
- Yes
- Privacy on Venice
- Anonymized — prompts not stored
- Available on Venice since
- Jun 2026
API
Call it from your code
Venice exposes this model through the REST API. Queue a generation with the model id.
curl https://api.venice.ai/api/v1/video/queue \
-H "Authorization: Bearer $VENICE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "wan-2-7-uncensored-text-to-video",
"prompt": "Aerial drone shot over a misty mountain valley at golden hour"
}'
# Use the returned queue_id with https://api.venice.ai/api/v1/video/retrieve.
# Call /video/complete after downloading if needed.Pricing
What it costs on Venice
Pay per clip on Venice — price scales with resolution and duration (5s–15s), from $0.57.
New Venice accounts include a free daily allowance and 500 welcome credits — no credit card required.
Alternatives
How it compares
| Model | Best for | Max resolution | Audio support | Open weights | Price (Venice) |
|---|---|---|---|---|---|
| Wan 2.7 Uncensored | Uncensored, open-weight video generation with native audio. | 1080p | Yes (Native) | Yes | from $0.57 |
| Wan 2.7 | The standard, moderated version of Alibaba's 27B video model. | 1080p | Yes (Native) | Yes | from $0.55 |
| Seedance 2.0 | ByteDance's proprietary rival; strong on face fidelity and control. | 1080p | Yes | No | from $0.35 |
| Kling O3 Pro | Kuaishou's cinematic video generator; proprietary. | 1080p | Yes | No | from $0.46 |
Uncensored, open-weight video generation with native audio.
Use cases
What it is good for
- 01Creating uncensored cinematic video clips and narrative sequences.
- 02Generating high-fidelity social media content (9:16) with native synchronized audio.
- 03Prototyping video concepts and character-consistent scenes for indie filmmakers.
- 04Developing marketing assets and video ads without restrictive platform filters.
Prompting
Getting better results
Describe both the visual action and the desired audio/sound effects to leverage native audio conditioning.
Specify the aspect ratio (16:9, 9:16, or 1:1) to match your target platform.
Keep prompts descriptive but focused on a single coherent motion sequence for the best physical realism.
Version history
Predecessor video model praised for open weights.
Improved open-weight video generation.
Current — flagship 27B model with native audio and uncensored access.
FAQ
Frequently asked questions
Wan 2.7 Uncensored is an advanced 27-billion-parameter text-to-video model developed by Alibaba's Tongyi Lab. This uncensored version hosted on Venice removes restrictive corporate filters, allowing creators to generate high-fidelity video clips with native audio synchronization and complete creative freedom.
On Venice, pricing is pay-per-clip and scales with resolution and duration. It starts from $0.57 for a 720p, 5-second clip, and goes up to $0.88 for a 1080p, 5-second clip. Clips can be generated in lengths of 5, 10, or 15 seconds.
Wan 2.7 is an open-weight model released under the Apache 2.0 license, meaning its weights are publicly available for self-hosting. On Venice, you can try it using your account's daily free allowances or pay-per-clip using Venice credits.
Yes. Wan 2.7 features native audio conditioning, meaning it generates synchronized sound effects and audio cues directly aligned with the on-screen motion in a single pass.
The standard Wan 2.7 model includes built-in moderation and safety filters imposed by its creators. The Uncensored version on Venice removes these restrictive guardrails, giving you unbiased, permissionless generation for creative sovereignty.
Venice processes your requests under an anonymized privacy tier. Your prompts are forwarded to the inference provider without any identifying personal data, and Venice does not store, profile, or train on your prompts.
It supports 1080p and 720p resolutions, with selectable aspect ratios of 16:9 (widescreen), 9:16 (vertical), and 1:1 (square), in clip lengths of 5, 10, or 15 seconds.
Run Wan 2.7 Uncensored privately
No prompt logging. No data used for training.