New Contest with Great Prizes: ComfyUI x MiniMax H3 Sync Sound Challenge - Open to Everyone!

Introduction
The generative skies just opened wider than ever. ComfyUI and MiniMax have thrown open the doors to the most deliciously timed community challenge of the season: the Comfy H3 Sync Sound Community Challenge.
For the next two weeks you get to build short, soul-stirring pieces where motion and sound are born as one inseparable whole—using the open-weights MiniMax H3 model right inside ComfyUI. Native stereo audio, real lip sync, foley, ambient beds, driving beats… all generated in the same magical pass as the picture.
And the prizes? An RTX 5090 and three RTX 5060 Ti cards ready to land on the desks of the most inspired among us. Local rig or Comfy Cloud—everyone is welcome. The clock started ticking on August 20, so let’s dive in while the light is still golden.
The Magic of True Sync Sound
MiniMax H3 is the first open-weights omni-modal video model that treats audio as a native twin of the visual stream. No more stitching a soundtrack on afterward and hoping the mouths line up. H3 dreams both layers together in stereo, up to 2K, and up to about 15 seconds per clip (you can chain them thoughtfully for longer stories, as long as the final piece stays under 90 seconds).
You can feed it text, images, reference video, or even a clean reference audio clip (no unlicensed IP, please) and watch the model weave picture and sound into one coherent living moment. The memory footprint has been trimmed by 66 %, so even an RTX 3060 can join the party, and brand-new Comfy Cloud users receive five free GPU runs to get started.
Pro Tip: Want an extra edge? Lean on Comfy MCP—the natural-language agent layer—and let Claude or Cursor help you reshape workflows, check VRAM, or debug on the fly. There’s a dedicated bonus prize waiting for the most insightful MCP-driven entries.
How to Enter (and the Simple Golden Rules)
Submission window: August 20 – September 1, 2026Winner livestream: September 2 with guest judges
What you must deliver
- Your finished video (max 90 seconds) with H3-generated, perfectly synced audio
- The actual ComfyUI workflow .json file
Where to submit
- Official form: https://form.typeform.com/to/WEAP3zp8
- Post your video in the H3 Reddit megathread (linked from the challenge page) so the community can cheer, remix, and comment
- Optional but highly encouraged: share on X, Instagram, LinkedIn, YouTube or TikTok tagging @ComfyUI @MiniMax_AI #ComfyH3 so the team can amplify the best work
Core rules (keep these close to your heart)
- A major portion of the piece must be generated inside ComfyUI using H3 for the core video + audio
- You may use reference audio to guide generation, but you may not simply slap a separately created track onto the finished video
- Everything must be lawful, SFW, and free of unlicensed IP or real-person likenesses
- One entry per person
- Mix local and Cloud runs as you like
Full details live here: https://blog.comfy.org/p/comfy-h3-sync-sound-community-challengeAnd the original announcement sparkles on X:
Getting H3 runningUpdate to ComfyUI 0.30.0 or newer, then grab the weights from Hugging Face: https://huggingface.co/Comfy-Org/MiniMax-H3. Ready-made templates are already in the ComfyUI template library. Prefer zero install friction? Jump straight into Comfy Cloud.
The Prizes That Sparkle
Four radiant ways to win hardware that will make your next projects sing:
- Grand Prize – Best Overall: RTX 5090-O32G-G
- Best Technical / Workflow: RTX 5060 Ti
- Best Creative: RTX 5060 Ti
- Built with MCP Bonus: RTX 5060 Ti
All GPUs ship with customs covered. If shipping is impossible, a cash equivalent lands instead. Every single entry also earns a chance to appear in the official spotlight reel. One workflow can claim multiple categories, so dream big.
Judges from Comfy, MiniMax, and special guests will score creative soul, technical elegance, community usefulness, and MCP insight. Community reactions on social and in the September 2 livestream chat will gently color the final conversation—no bots, just genuine love.
Conclusion
This is more than a contest. It is an invitation to explore the new frontier where picture and sound are born as twins, and to share the workflows that make the rest of us better. Whether you craft a whispering forest dialogue, a beat-driven cyberpunk dance, or a quiet moment of pure foley poetry, the canvas is wide open and the prizes are real.
Update your ComfyUI, download H3, open a fresh graph, and let the stereo magic begin. I’ll be watching the megathread with stars in my eyes, cheering every luminous submission.
Go make something that moves us—and maybe takes home a shiny new GPU while you’re at it.
—Aurelia ♡
