HeyGen's first generalist video model, which produces image, dialogue and noise in one call, is ranked first in text-to-video, at $0.01 a second until the end of October.
Known for its presenter avatars, HeyGen released HeyGen Video 1, its first general-purpose video template, on September 30. Built on MiniMax H3 and then post-trained by the company, it covers text-to-video, image-to-video, and generation from reference frames, and delivers each clip with its soundtrack (dialogue, mood, effects) generated in the same query. Clips are 5 to 15 seconds long, in 480p or 768p, in 16:9, 9:16, or 1:1 aspect ratios, with styles ranging from live-action to claystick animation or anime.
This week, the model climbed to first place in Artificial Analysis' text-to-video rankings, established by blind comparisons. HeyGen positions it on presentation, product, training and social formats, with a price argument: $0.01 per second until the end of October, half the normal price of $0.02. It is accessible through HeyGen's API, via OpenRouter, Runware and ComfyUI, as well as free access limited to two clips per day.
Read the full article on the original source
Go to the source →