Wait Which Model?
← Back to Video models
HeyGenNotable

HeyGen Video 1

Released Sep 30, 2026

HeyGen's first general-purpose video model, built on MiniMax H3 and post-trained by HeyGen. It generates a whole scene with picture and sound in one call, from a text prompt, a first-frame image or reference files.

Price per second
—
Max clip length
30 s
Max resolution
2K
Native audio
Yes
AA arena Elo
1157
LMArena Elo
—
Capabilities
text to video, image to video
Licence
Proprietary
Strengths & weaknesses

Strengths

  • No. 4 on Artificial Analysis' text-to-video leaderboard with audio (Elo 1157)
  • Dialogue, ambience and sound effects generated with the picture (32 kHz stereo AAC)
  • Up to 12 references (9 images, 3 videos, 3 audio clips) in reference-to-video mode
  • Output from 480p to 2K

Weaknesses

  • API-only at launch; web app access announced but not dated
  • Reference-to-video limited to 15 seconds
  • HeyGen's own pages disagree on the per-second price

releaseDate is HeyGen's own stated launch date ('launched on September 30, 2026', heygen.com/blog/best-video-models). The September 2026 API changelog also carries the first heygen-video-1 reference, while AA lists release as 'Oct 2026'. Built on MiniMax H3 and post-trained by HeyGen (developer docs and changelog). Duration conflict: the model docs page says text- and image-to-video run 5 to 30 s 'with compatible settings' and reference-to-video 5 to 15 s; the Create reference page, the September changelog and the launch blog say 5 to 15 s. 30 is used. Resolution: 480p and 768p at launch (default 768p); October 2026 changelog entries add 1080p and 2K. Pricing conflict, all on HeyGen pages: the API pricing page shows 50%-off launch rates of $0.01 (480p), $0.018 (768p) and $0.05 (1080p/2K) per second for text-to-video, list double; the cheapest-video-models blog gives $0.02 (480p) and $0.03 (768p) per second with audio, 1080p/2K at three times 768p, half off through Oct 31, 2026; the model page says '$0.01 per second' launch price through October. perSecondUsd is therefore null. AA lists $2.40/min (1080p basis). No weights found on Hugging Face and HeyGen describes it only as a hosted API, so openWeights false. Not on LMArena's video board. Notable by clause 1: AA-Video-T2V v2.1 #4 (2026-10-27). HeyGen's '#1 / 9.3/10' rankings on its blog are self-assessed and not used.

Links

API model strings

  • heygenheygen-video-1