HeyGen has introduced HeyGen Video, a general-purpose video-generation model designed to create a scene's visuals and audio together. In its announcement published on October 21, 2026, the company described a system that can generate clips from text prompts, images or reference videos. The model is intended to produce the picture, dialogue and sound effects in a single generation process rather than assembling them from separate tools.
According to HeyGen, the system supports clips lasting between 5 and 15 seconds, with output resolutions reaching 2K. The company made the model available through its API and described a launch price of $0.01 per second through October. Its announcement also highlighted the ability to work across visual styles, including live action and stop-motion.
Generating visuals and sound together may help creators produce short videos for educational explainers, product demonstrations and social media without coordinating several separate generation stages. For learners exploring generative AI, the release also shows how video models are moving beyond silent clips toward scenes with accompanying audio.
HeyGen's announcement said the model was available through its API. The company's product materials describe text, image and reference-based video generation, giving developers several ways to supply the starting material for a clip.