Creatify Boreal model_version on AI Avatar vs Sume quality tiers

Creatify's API changelog shows a boreal model_version option on its AI Avatar v1 and v2 endpoints, plus URL to Video and Custom Template, added in September 2026. Sume exposes a simpler switch: quality with standard, plus (the default) and max. Both let you trade cost against fidelity per request; the difference is in the limits that surround them.

Creatify's limits as published

The changelog lists AI Avatar v1 with audio up to 60 seconds and text up to 800 characters, and AI Avatar v2 with 800 characters per input_text. Its Boreal-H3 endpoint is listed at 5 to 15 seconds and 10 credits per 30 second video.

Sume's equivalent

quality on POST /v1/avatar-1.0/talking-video is the model-class switch. Per second with no product image: standard $0.184, plus $0.245, max $0.55. Estimated duration must be 4 to 60 seconds, and resolution is 720p.

See the avatar video docs. Override quality at generate-video from a preview without making a new one.

Practical difference

With a character cap per input you chunk text; with a duration window you size the script. Sume asks for exactly one of script or video_inputs, and scenes carry a duration for silence.

Check Creatify's changelog for current numbers; limits change often.