HeyGen Launches Its First General-Purpose Video Model at $0.01/s, Claiming to Beat Veo 3.1
HeyGen · x · 2026-10-01
HeyGen has released HeyGen Video, its first general-purpose video model, built on MiniMax H3 and post-trained by HeyGen. It handles text-to-video, image-to-video and reference-to-video in one model, with sound (dialogue, ambience, effects) generated in the same call.
Key facts:
- Pricing: $0.01/s through October (50% off the standard $0.02); live on the HeyGen API, OpenRouter, Runware and ComfyUI
- Internal eval (Artificial Analysis Arena set, 4,800 votes): normalized Elo of 1000, above Seedance 2.0 (955), H3 Max (952), Kling 3.0 Pro (857) and Veo 3.1 (744)
- Cost: $0.03/s vs Kling 3.0 Pro $0.168/s, Seedance 2.0 $0.303/s and Veo 3.1 $0.40/s
- Speed: 3.7s inference for a 10s I2V clip, faster than H3 Max Turbo (4.1s) and H3 Max (8.3s)
- Blind preference: best-in-class on image fidelity (65%) and prompt following (62%), weaker on consistency and detail
Aimed at business video: product demos, training, onboarding and property walkthroughs.
Related event: HeyGen Launches Video Generation Model at $0.01 per Second(2 posts)→
More from Multimodal
- Generate Video by Clicking and Dragging: H3 Max Turbo on fal Runs Faster Than Real Time — adamho · 2026-10-01
- Indie dev's AI-built dream game, day 17: a harpy unit in just over a day — majidmanzarpour · 2026-10-01
- Native 1080p output: spend expensive compute only after creative decisions are locked — mhdfaran · 2026-10-01
- Seedance 2.5 workflow: draft at 480p, then re-render your pick at native 1080p — mhdfaran · 2026-10-01
- Generating Video with Just a Click and Drag: New Interactive Demo — adamho · 2026-10-01
- 14 years after AlexNet hit ImageNet, a dev recreates the moment with a $0 open-source pipeline on one RTX 3090 — onil_gova · 2026-10-01