For Malaysian creators and small teams, the cost of a short marketing video has long been the gap between an idea and a finished clip. HeyGen wants to move that number. The company has released HeyGen Video 1.0, its first general-purpose video model, and it generates clips, with sound, from a single prompt.
Contents
What HeyGen Video 1.0 does
One model handles three jobs that used to need separate tools. You can start from a line of text (text to video), from a single opening frame (image to video), or from your own reference images (reference to video). Every clip comes back with sound generated in the same request, covering dialogue, ambience and effects, so there is no separate audio pass. HeyGen is showing the model across several looks, from live action to clay, anime and illustrated styles, in its launch catalogue of sample clips.
The price is the headline
Pricing starts at $0.01 per second through October, which HeyGen says is half its standard rate of $0.02. That works out to roughly RM0.05 a second at current exchange rates, so a 30 second clip sits at about RM1.40 before any editing. The model is live now through the HeyGen API and on OpenRouter, Runware and ComfyUI, so developers and studios anywhere, Malaysia included, can call it without waiting for a regional rollout.
Why it matters for Malaysian teams
HeyGen is aiming this at everyday business work rather than showreels: product demos, training, onboarding and property walkthroughs. That is the kind of video a local agency or an online seller makes constantly and usually pays a crew to shoot. The company says that in blind side by side tests the model performs alongside the top video models at a fraction of their price per second. That is HeyGen's own claim, so it is worth running on your own brief before you budget around it, but the pricing alone changes the maths for short, repeatable clips. The model is built on MiniMax H3 and post-trained by HeyGen.
It also lands in a year when generative tools keep closing the gap with real production. We have already looked at AI that can clone a voice in ten seconds, and at free apps that turn a phone into a proper video camera. A text to video model with built-in sound is the next piece, and it sits alongside a real camera rather than replacing the reasons to pick one up.
The takeaway
AI video has spent two years looking good in demos. A price of one US cent a second, with sound in the same call, is what turns it into something a small Malaysian team can actually put in a monthly budget. The real test is no longer whether a clip shines in a launch reel, but whether it holds up on your own product, and that is now a test anyone can run for the price of a coffee.
