Wan 2.7
Alibaba's Wan 2.7 is a unified image and video generation model with open weights, native audio, and instruction-based editing — built for production workflows.
Overview
What is Wan 2.7
Wan 2.7 is Alibaba's open-weight text-to-image and image-to-video model, released on April 1, 2026. It supports high-resolution image generation, 1080p video up to 15 seconds, native audio co-generation, and instruction-based video editing — all under an Apache 2.0 license for commercial use.
Running it privately on Venice
On Venice, Wan 2.7 runs with full prompt anonymity — your inputs are not stored or profiled. This enables private, permissionless creation using one of the few open-weight models capable of both high-quality image and video synthesis. You gain sovereign control over visual content without Big Tech surveillance.
Assessment
Strengths and limitations
- Open-weight under Apache 2.0: you can self-host, fine-tune, or integrate without licensing risk.
- Unified model for both image and video generation, reducing pipeline fragmentation.
- Supports instruction-based video editing: modify existing clips with natural language commands.
- Native audio co-generation with phoneme-level lip sync and voice cloning capabilities.
- Flexible input modes: text, image, or reference video to drive generation.
- Max clip length capped at 15 seconds: shorter than some cinematic use cases require.
- 1080p resolution limit; no native 4K output like in Kling 3.0 or Sora 2.
- No tool use, web search, or reasoning capabilities — strictly multimodal generation.
- Still requires prompt engineering for consistent character or object persistence.
Samples
Sample outputs
Generated on Venice with our standard prompt suite — the same prompts we run through every model of this type, so you can judge it like-for-like.
A retro travel poster with the bold headline "VENICE" in large condensed serif type, sunset color palette, clean layout
Photorealistic close-up portrait of a weathered fisherman at golden hour, 85mm lens, shallow depth of field, natural skin texture
A small red cube balanced on top of a large glossy blue sphere, with a green cone to the right, plain light-grey studio background
Cozy watercolor illustration of a hillside village in autumn, warm tones, soft paper texture
Capabilities
What it supports
- Text to image
- Image to image
Specifications
Datasheet
- Maker
- Alibaba
- Released
- April 1, 2026
- Modality
- Text-to-image, image-to-video, reference-to-video
- Max resolution
- 1080p
- Clip length
- Up to 15 seconds
- Open weights
- Yes — Apache 2.0
- Aspect ratios
- 1:1, 3:2, 16:9, 21:9, 9:16, 2:3, 3:4, 4:5
- Prompt limit
- 3,000 chars
- Privacy on Venice
- Anonymized — prompts not stored
- Available on Venice since
- Apr 2026
- License
- Apache 2.0
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/image/generate \
-H "Authorization: Bearer $VENICE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "wan-2-7-text-to-image",
"prompt": "A serene mountain lake at dawn, photorealistic"
}' --output image.pngPricing
What it costs on Venice
Flat per-image pricing on Venice: $0.04 per generation.
New Venice accounts include a free daily allowance and 500 welcome credits — no credit card required.
Alternatives
How it compares
| Model | Max resolution | Strongest at | Open weights | Price (Venice) |
|---|---|---|---|---|
| Wan 2.7 | 1080p | Flexible video workflows | No | $0.04 / image |
| Grok Imagine High Quality (SOTA) | 4K | Cinematic quality | No | from $0.06 / image |
| Chroma | 1080p | Speed & open weights | Yes | $0.01 / image |
| Lustify v8 | 1080p | Adult content generation | No | $0.01 / image |
Open-weight advantage with editing and reference inputs.
Use cases
What it is good for
- 01Branded short-form video with consistent voice and visuals across clips.
- 02Social media content where native audio and lip sync improve engagement.
- 03Internal creative tools that require self-hosting or on-premise deployment.
- 04Video editing workflows driven by natural language instructions.
- 05Multimodal storytelling using image, text, and audio in a single pipeline.
Prompting
Getting better results
Use reference images or videos to lock in style, motion, or composition.
Specify audio requirements clearly: 'voiceover in Spanish', 'background music: ambient synth'.
For editing, describe the change precisely: 'zoom in on the character's face', 'replace the background with a forest'.
Leverage aspect ratio support for platform-specific formats: 9:16 for TikTok, 16:9 for YouTube.
Version history
Predecessor with basic video generation.
Current — adds open weights, audio, editing, reference inputs.
FAQ
Frequently asked questions
Wan 2.7 is Alibaba's open-weight multimodal model for text-to-image and image-to-video generation, released on April 1, 2026. It supports 1080p video up to 15 seconds, native audio, voice cloning, and instruction-based editing under an Apache 2.0 license.
On Venice, Wan 2.7 costs $0.04 per image generation. Upscaling is $0.02 for 2× and $0.08 for 4×. Video generation pricing is not included in this tier.
Wan 2.7 is not free, but it is open-weight under the Apache 2.0 license, meaning you can download, modify, and commercially use the model weights without restriction.
Yes. Wan 2.7 supports text-to-video, image-to-video, and reference-to-video generation with up to 15 seconds of 1080p output, including native audio and lip sync.
Yes. Wan 2.7 supports instruction-based video editing — you can upload a clip and apply natural language commands like 'zoom in' or 'change background' to modify it.
No. Wan 2.7 is a pure generation model and does not support tool use, web search, or reasoning. It focuses on high-quality image and video synthesis from text, image, or reference inputs.
Wan 2.7 excels in open-weight flexibility, video editing, and native audio, while GPT Image 2 leads in in-image text accuracy and 4K resolution. Choose Wan 2.7 for editable, audio-rich video; GPT Image 2 for precise, text-heavy visuals.
Wan 2.7 supports up to 1080p resolution for both images and videos, with aspect ratios including 1:1, 16:9, 9:16, and others optimized for social and cinematic formats.
Wan 2.7 is not uncensored — it includes standard content filters. However, its open-weight nature allows developers to deploy modified versions without filters in permissionless environments.
Run Wan 2.7 privately
No prompt logging. No data used for training.