MiniMax Just Dropped a Open Source "Seedance Killer"

From the creator

Today's video is sponsored by ElevenLabs! Try ElevenAgents and start chatting with your characters — plus grab 10,000 free credits courtesy of ElevenLabs right here: https://try.elevenlabs.io/w5183uo8pydc This is a surprise: MiniMax is BACK. After nine months of silence since Hailuo 2.3, they’ve launched Hailuo 03 (aka MiniMax H3) — a fully multimodal AI video model with native audio, up to 15-second generations at 2K, and an Omni mode that takes up to 12 image references plus video and audio inputs. Is it the Seedance killer people are claiming? I ran it through the full gauntlet to find out. In this video, we test Hailuo H3 across cinematic text-to-video (including dialogue, acting range, and multilingual support), image-to-video against the old 2.3 model, first frame/last frame, video extensions, and multi-reference character team-ups — plus what I’m seeing on pricing, which looks like it comes in around 50% under Seedance 2.0. Key takeaways: dialogue and lip sync are shockingly good, fight choreography trades speed for stability (no morphing, no teleporting), prompts cap out at a generous 7,000 characters, and the Omni model takes your storyboards very seriously — maybe more seriously than Seedance does. Note: H3 is now Open Source (couldn't say that when the video first came out) and is available over on Hugging Face: https://huggingface.co/MiniMaxAI/MiniMax-H3 CHAPTERS 00:00 Intro: MiniMax Is Back 00:49 What's New in Hailuo 03 01:44 Dialogue & Acting Test (Twin Peaks Diner) 02:21 Cinematic Text-to-Video Tests 03:02 The Mob Movie Test 03:47 Multilingual Dialogue (Samurai Test) 04:39 Fight Scenes + the 7,000-Character Prompt Limit 06:10 Image-to-Video: Nine Months of Progress 07:03 Flamethrower Girl Returns 07:50 First Frame / Last Frame 08:30 The Omni Model: Multi-Reference 09:37 ElevenLabs Sponsor Segment 13:34 Sniper Girl One-Shot 14:13 Video References & Extensions 15:18 Sniper + Flamethrower Team-Up 16:02 "Reference Maxing" & Storyboards 17:25 Pricing Ballpark 18:23 Final Thoughts

Choose to Build with AI
Matched to Minimax Algorithm

AI Maker Residence 3

The third AI workshop taught by our legendary teacher, Nick Sarafa. This is a full-day hands-on training workshop for purposeful co-creation with AI using Claude Code. Imagine having access to hundreds of billions of dollars of computing power and knowing exactly how to make it work for you through the power of super intelligence.

◆ Fri 09 Oct 2026 ◆ KOKO Cafe, London ◆ With Nick Sarafa
AI Maker Residence 3
Live event
AI Maker Residence 3
Fri 09 Oct 2026

More like this

Running one yourself?

List your AI event,
wherever it is.

A meetup, a workshop, a hackathon, a conference. Any city, or online. Tell us about it and it lands in front of people already learning this stuff.