Noesion Radarnoesion.ai
Release2026-10-04

FlashML released FreeVideo, a local inference engine for MiniMax H3. It runs on consumer GPUs with as little as 8 GB VRAM and 16 GB RAM. The engine adapts its acceleration path to the available hardware. It supports text, image, video and audio inputs. FreeVideo integrates with ComfyUI and offers a Windows launcher for setup. It also supports command-line use on Linux. The project is open-source under Apache 2.0. Model weights carry separate MiniMax H3 Community License restrictions.

What changed

Running MiniMax H3 required more VRAM and RAM than many consumer GPUs provide.

What it unlocks

Generating videos locally on consumer GPUs with MiniMax H3.

  • 8 GB VRAM minimum
  • 16 GB RAM minimum

What you need to act on it

  • a consumer GPU
  • ComfyUI or command-line setup

Sources

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.