HappycapyGuide

By Connie · This article contains affiliate links. We may earn a commission at no extra cost to you if you sign up through our links.

AI Video Tools

Netflix Just Released Its First Public AI Model — and It Removes Objects From Video With Physics Awareness

Netflix open-sourced a physics-aware video AI on April 2, 2026 — the first public model from a major streaming company. It removes objects and their causal effects from video.

April 2, 2026 · 7 min read · By Connie

TL;DR

Netflix released void-model on Hugging Face on April 2, 2026 — a 5-billion-parameter video AI that removes objects from footage and erases their physical effects (shadows, reflections, collisions). Built on CogVideoX-Fun-V1.5-5b. Supports 197 frames at 384×672. Requires 40GB+ VRAM. Available on GitHub. Research-oriented, not production-ready. First public AI model Netflix has ever released.

5B
parameters
197
max frames
40GB+
VRAM required
1st
Netflix public AI model

What Netflix Actually Released

Netflix open-sourced void-model on April 2, 2026 via Hugging Face and GitHub. The model is formally named netflix/void-model and is built on a fine-tuned version of CogVideoX-Fun-V1.5-5b — a 5-billion-parameter video generation model developed by Alibaba's research team.

The model's core function is object removal from video footage. You mark a region using a "quadmask" input — a bounding box or segmentation mask — and void-model removes the object in that region across all frames of the video clip. It supports clips up to 197 frames at 384×672 pixel resolution.

What makes this different from existing video inpainting tools is the physics-awareness. Void-model does not simply erase pixels and fill the gap with background texture. It also removes the causal downstream effects of the erased object — shadows the object cast, reflections it produced in shiny surfaces, and physical interactions it caused with other objects in the scene.

What "physics-aware" actually means: If a removed object knocked over a cup, the cup rights itself in the output. If the object cast a shadow on a wall, the shadow disappears. If it created a ripple in water, the ripple is removed. Standard inpainting fills the object's silhouette but leaves all of these secondary effects intact — making the edit visible.

Technical Specifications

Void-model is derived from CogVideoX-Fun-V1.5-5b, a video generation model trained by Alibaba's CogVideoX team and released under Apache 2.0. Netflix fine-tuned this model specifically for the object-removal task using its own internal VFX datasets — which almost certainly include footage from Netflix productions.

SpecificationValue
Base modelCogVideoX-Fun-V1.5-5b (Alibaba)
Parameters5 billion
Max resolution384×672 pixels
Max frames197
VRAM required40GB+ (A100 or equivalent)
Input formatQuadmask (bounding box or segmentation)
LicenseResearch (not commercial production)
Available onHugging Face, GitHub
Release dateApril 2, 2026

Why This Is Significant

Netflix has been one of the largest private consumers of AI technology for over a decade — its recommendation engine, thumbnail optimization system, and content delivery infrastructure are all AI-powered. But the company has never released a public AI model until now.

The decision to open-source void-model marks a shift. Netflix is not releasing this as a product — it explicitly describes void-model as "research-oriented, not production-ready." But the release positions Netflix in the AI research community and likely serves as recruiting signal for ML engineers who want to work with proprietary production VFX data.

For the VFX and post-production industry, the physics-aware removal capability addresses a real bottleneck. Current AI inpainting tools (Adobe Firefly, Runway Inpaint, Topaz Video AI) can fill the pixel region an object occupied, but they do not model the object's physical interactions. VFX artists must manually clean up secondary effects, which can take hours per shot. Void-model automates that second pass.

Need AI tools for video and content work?
Happycapy gives you access to the latest AI models for video analysis, content creation, and research — all in one workspace with no setup required.
Try Happycapy Free →

Who Can Actually Use Void-Model Right Now

The 40GB+ VRAM requirement puts void-model out of reach for consumer hardware. A standard gaming GPU (RTX 4090) has 24GB VRAM — not enough. Running the model requires professional data center GPUs like the NVIDIA A100 (40GB or 80GB variant) or H100.

User TypeCan Run Void-Model?How
VFX studio / production houseYesOn-premises A100/H100 cluster or cloud GPU
Individual ML researcherYes (via cloud)Google Colab A100, Lambda Labs, RunPod
Game developer / indie studioNoRTX 4090 = 24GB, below requirement
Consumer / creatorNoConsumer hardware insufficient
Enterprise AI teamYesAWS p4d, Azure NC A100, GCP A100 VMs

How Void-Model Compares to Existing Video AI Tools

ToolTypePhysics-AwareAccessHardware Need
Netflix void-modelObject removalYesOpen source (research)40GB+ VRAM
Adobe Firefly VideoInpainting / generativeNoAdobe subscriptionCloud (browser)
Runway Gen-3 InpaintInpaintingNoRunway subscriptionCloud (browser)
Topaz Video AIEnhancement / removalNoOne-time purchase4GB+ VRAM
DaVinci Resolve Magic MaskMasking + removalNoFree / Studio licenseGPU varies

Frequently Asked Questions

What is Netflix void-model?
Netflix void-model is an open-source 5-billion-parameter video AI released on April 2, 2026 that removes objects from video while also erasing their physical effects — shadows, reflections, and collision interactions. It is built on CogVideoX-Fun-V1.5-5b and requires 40GB+ VRAM. Available on Hugging Face and GitHub.
What makes void-model different from regular video inpainting?
Standard inpainting tools erase pixels and fill the gap with background content. Void-model is physics-aware — it also removes downstream effects: shadows the object cast, reflections it created, and physical interactions it caused with other objects. This eliminates the secondary cleanup step that currently requires manual VFX work.
Can I run void-model on my consumer GPU?
No. Void-model requires 40GB+ VRAM, which exceeds consumer GPUs like the RTX 4090 (24GB). You would need a professional data center GPU (NVIDIA A100 or H100) or access to cloud GPU services (Google Colab A100, Lambda Labs, RunPod) to run inference.
Is void-model production-ready?
Netflix explicitly describes void-model as "research-oriented, not production-ready." It is intended for experimentation, VFX research, and academic use — not for direct integration into commercial production workflows without additional engineering work.
Sources
  • Netflix / Hugging Face: netflix/void-model model card (April 2, 2026)
  • GitHub: netflix/void-model repository — inference scripts and technical documentation
  • Alibaba CogVideoX-Fun-V1.5-5b — base model documentation
  • TechCrunch: "Netflix releases its first open-source AI model" (April 2, 2026)

Related reading: DeepMind AlphaEvolve · AI agent use cases 2026 · How to use AI for video editing

Work with the latest AI models without the setup
Happycapy gives you instant access to 150+ AI models — Claude, GPT-4o, Gemini, and more — in a single workspace. No API keys, no VRAM requirements, no configuration.
Try Happycapy Free →
SharePost on XLinkedIn
Was this helpful?

Get the best AI tools tips — weekly

Honest reviews, tutorials, and Happycapy tips. No spam.

Comments