HunyuanVideo

Self-Hostable

Category: video • Released 2024

Text-to-video research model with high-memory GPU requirements

The official README reports peak GPU memory near 45 GB for 544×960×129 frames and 60 GB for 720×1280×129 frames, with NVIDIA CUDA/Linux tested and an 80 GB GPU recommended. This is not a typical consumer GPU workflow.

Pricing Structure & Model AccessTencent Hunyuan Community License Agreement; not a standard MIT/Apache license.

freeOfficial Pricing

Free Access QuotaLocal weights; no hosted generation service included.

Free Tier LimitsOfficial documentation lists 45–60 GB peak memory for documented configurations; 80 GB GPU recommended.

Requirements & Verification

No Account Required (Instant)Verified: 2026-10-23

Open weights · community license · 45–60 GB VRAM

License: Tencent Hunyuan Community License; territory and use restrictions apply.

Source repository Model weights

Core Capabilities & Features

Text-to-video model

Official inference code

Community-license terms

Local Linux/CUDA workflow

Who It's Best Suited For

Research teams with large NVIDIA GPU memory and license eligibility.

Execution Environment

Linux with NVIDIA CUDA; local PyTorch inference.Hardware / setup: 45 GB peak at 544×960; 60 GB at 720×1280; 80 GB recommended (official README).

Tags & Keywords

#video #open-weights #hunyuan #text-to-video #tencent #gpu

Side-by-Side Comparison

Evaluate licensing, rate limits, and pricing models directly against similar platforms.

Compare HunyuanVideo

Related alternatives to HunyuanVideo

These options overlap with parts of this workflow. Check the fit note, setup, and license before switching.

Runway AI Video

video

Freemium

AI video generation and editing tools for image-to-video and creative production workflows.

Text-to-videoImage-to-videoCamera control

Runway plan credits and export conditions change over time. Check the official pricing page before estimating the cost of a project.

Explore

Website

Kling AI

video

Freemium

AI video generation tools for text-to-video and image-to-video workflows.

Text-to-videoImage-to-videoCamera motion controls

Check Kling's current pricing and in-product account page for available free credits, export conditions, and paid plan limits.

Explore

Website

Pika

video

Credits

Idea-to-video platform with playful physical Pikaffects and granular motion control

Pikaffects physics engineLip sync audio matchSound effects auto-generation

Trial credits on initial signup

Explore

Website

Luma Dream Machine

video

Freemium

High-speed, physics-accurate generative video with photorealistic camera motion

Physics-accurate simulationCinematic camera movesKeyframe continuation

Free plan with 30 monthly generations

Explore

Website

Similar video AI Platforms

View all video tools →

Runway AI Video

video

Freemium

AI video generation and editing tools for image-to-video and creative production workflows.

Text-to-videoImage-to-videoCamera control

Runway plan credits and export conditions change over time. Check the official pricing page before estimating the cost of a project.

Explore

Website

Hailuo AI (MiniMax)

video

Freemium

Flagship video-01 foundation model with fluid physical dynamics and cinematic realism

Fluid physical dynamicsFacial micro-expressionsText-to-video

Free tier with daily generation allowance

Explore

Website

Wan 2.1 (Alibaba)

video

Open Weights

Apache 2.0-licensed open-weight video generation model with local and hosted execution options.

Apache 2.0 model weightsText-to-videoImage-to-video

The model weights in the official repository are licensed under Apache 2.0. Local compute and hosted inference are separate costs.

Explore

Source

Related Workflows

Step-by-Step Guides for HunyuanVideo

Browse all guides →

video creation

A Practical Cinematic AI Video Workflow with FLUX and Runway

A reproducible workflow for planning a shot, generating a reference frame, testing image-to-video motion, and finishing the edit.

#video#flux#runway

4 min read

Read Guide