Zen LM
Models

Video

Zen video generation models for professional text-to-video and image-to-video synthesis.

Video

The Zen Video family delivers professional-grade video synthesis from text and image inputs. Built on diffusion-transformer architectures, these models enable high-resolution video generation for creative, media, and design workflows.

Models

ModelParamsContextHFPaper
Zen Director5Bweightspaper
Zen Video13Bweightspaper
Zen Video I2V13Bweightspaper

Quick Start

Text-to-Video with Zen Director

The Zen Director pipeline provides the most efficient text and image-to-video generation:

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("zenlm/zen-director")
tokenizer = AutoTokenizer.from_pretrained("zenlm/zen-director")

from zen_director import ZenDirectorPipeline

pipeline = ZenDirectorPipeline.from_pretrained("zenlm/zen-director")
video = pipeline(
    prompt="A cinematic shot of a sunset over mountains",
    num_frames=120,
    fps=24,
    resolution=(1280, 720)
)
video.save("output.mp4")

Using the Zen API

For production deployments, use the OpenAI-compatible Zen API endpoint:

from openai import OpenAI

client = OpenAI(
    base_url='https://api.hanzo.ai/v1',
    api_key='your-api-key',
)

response = client.images.generate(
    model='zen-video',
    prompt='A drone flying over a tropical coastline at golden hour',
    size='1280x720',
)
print(response.data[0].url)

Image-to-Video with Zen Video I2V

Animate a still image into fluid motion:

from zen_video_i2v import ZenVideoI2VPipeline
from PIL import Image

pipeline = ZenVideoI2VPipeline.from_pretrained("zenlm/zen-video-i2v")

image = Image.open("input_image.jpg")
video = pipeline(
    image=image,
    prompt="The camera slowly pans across the scene",
    num_frames=120,
    fps=24,
)
video.save("output.mp4")

Model Details

Zen Director is a compact 5B diffusion-transformer optimized for both text and image inputs, providing fast professional-grade synthesis.

Zen Video is the flagship 13B model delivering high-quality 720p text-to-video generation with precise prompt adherence.

Zen Video I2V specializes in image-to-video animation, converting still images into coherent, fluid video sequences with motion control.

All models support standard transformer inference with transformers and are optimized for Zen Engine deployment at 44K tokens/sec on Apple M3 Max hardware.

On this page