Model Variants

Choose the Right Model for Your Use Case

From lightweight 1.3B models running on consumer GPUs to 14B flagship models delivering cinematic quality.

Flagship

Wan 2.7 T2V-14B

14B480P / 720PVRAM: 16GB+

Flagship text-to-video model. Superior quality, handles complex scenes, supports both 480P and 720P output. Ideal for professional workflows.

Lightweight

Wan 2.7 T2V-1.3B

1.3B480PVRAM: 8GB

Lightweight model that runs on consumer GPUs. Generates 5-second 480P videos on RTX 4090 in ~4 minutes. Perfect for local deployment.

Wan 2.7 I2V-14B

14B480P / 720PVRAM: 16GB+

Image-to-video generation. Animates your still images with natural motion. Aspect ratio follows the input image automatically.

Wan 2.7 FLF2V-14B

14B720PVRAM: 16GB+

First-Last-Frame-to-Video — define the start and end frames, let the model generate the transition. Perfect for controlled sequences.

Editor

Wan 2.7 VACE-14B

14B480P / 720PVRAM: 16GB+

All-in-one video creation and editing. Supports R2V, V2V, MV2V tasks. The most versatile model in the Wan family.

Wan 2.7 VACE-1.3B

1.3B480PVRAM: 8GB

Compact video editor for consumer hardware. Great for quick edits, prototyping, and reference-guided generation.

Available via API

Access all Wan 2.7 models through Alibaba Cloud Dashscope API or fal.ai. No GPU required — pay per generation.

☁️

Alibaba Dashscope API

Official API from Alibaba Cloud. Stable, production-ready, supports all model variants.

Get API Key →

fal.ai

Fast, serverless inference. Pay per request. Ideal for prototyping and burst workloads.

Get API Key →