AI models

VidMachine uses leading AI video and image models so you get the best quality for your shorts and TikToks. Choose from multiple video models (including Google Veo 3.1, Veo 3.1 Lite, OpenAI Sora 2, xAI Grok, and Kling Avatar V2 for AI Influencer) and image models (including Google Nano Banana, OpenAI GPT Image 2, ByteDance Seedream, and Flux 2 Pro) when you create projects. Each model has different strengths in speed, style, and cost—explore the details below.

AI video models

Veo 3.1

Google's flagship video model with native audio, reference image guidance, and strong cinematic quality at 720p–1080p.

Veo 3.1 Lite

Google's cost-efficient Veo variant with native audio, image-to-video, and start/end frame control—ideal for high-volume shorts at 720p–1080p.

Sora 2

OpenAI's video and audio model with improved physics, controllability, and synchronized dialogue and sound effects.

Grok Video

xAI's image-to-video and text-to-video model with strong cinematic motion and competitive quality and speed.

WAN 2.5

Alibaba's image-to-video model with 720p–1080p output, audio sync, and multilingual prompt support.

WAN 2.7 I2V

Alibaba Wan 2.7 image-to-video with text-guided motion, optional first-and-last-frame control, clip continuation, and audio sync—on VidMachine we generate at 720p only.

Happy Horse 1.0

Alibaba’s Happy Horse 1.0 on Replicate: text-to-video or image-to-video at 720p, 3–15s, vertical (9:16) for text-to-video—integrated on VidMachine at 720p and 28 credits per second.

Seedance 1.5

ByteDance's audio-visual video model with native lip-sync, professional camera control, and 720p–1080p output.

Seedance 2.0

ByteDance's multimodal video model with native audio and reference images/video/audio, plus editing and extension workflows—available in VidMachine at 480p or 720p depending on your project settings.

Kling Avatar V2

Talking avatar video from each scene's starting frame and speech audio—lip sync and expressions follow your audio. AI Influencer projects only; VidMachine uses standard (std) mode at 11 credits per second.

AI image models

Nano Banana

Google's fast AI image model with strong quality and efficiency for everyday creative use.

Nano Banana 2

Google's latest image model (Gemini 3.1 Flash Image) with Pro-level quality, character consistency, and accurate text.

Grok Image

xAI's photorealistic image model with detailed textures, lighting, and up to 1024×1024 resolution.

Seedream 4.5

ByteDance's image model with exceptional text rendering, 4K output, and strong multi-image consistency.

Flux 2 Pro

Black Forest Labs' professional image model with multi-reference control, editing, and up to 4MP output.

WAN 2.7 Image

Alibaba's Wan 2.7 Image model for text-to-image and multi-image editing, supporting up to 2K output and coherent image sets.

GPT Image 2

OpenAI's image generation and editing model on Replicate—photorealism, strong text rendering, and precise edits; VidMachine uses medium quality and center-crops to 9:16 for shorts.