Read Seedance 2.5 and Wan 3.0 limits from GET /v1/videos/models

Call GET /v1/videos/models on Sume to read supported_durations, resolutions, references and generate_audio for seedance-2.5 and wan-3.0 before you submit.

4 min readSume
All posts

GET https://api.sume.com/v1/videos/models returns every video model with its supported durations, resolutions, aspect ratios, input references and generate_audio. Read it for seedance-2.5 and wan-3.0 before you hardcode a limit.

The docs say limits differ for each model, so a stored copy of this post's table can go stale.

Call

curl "https://api.sume.com/v1/videos/models" \
  -H "Authorization: Bearer $SUME_API_KEY"

Fields worth reading

Fields in each data[] entry (read 2026-10-07)
FieldUse
supported_durationswhole seconds the model accepts
supported_resolutionsfor example 480p, 720p, 1080p
supported_aspect_ratiosfor example 16:9, 9:16
supported_input_referencesimage_url, video_url, audio_url
generate_audiowhether an audio track is possible
pricing_skusprice information per SKU

What the docs say today

Validate the request client-side against this list to avoid a 400.

  • seedance-2.5: 4 to 30 s at 480p, 720p, 1080p.
  • wan-3.0: 2 to 30 s.
  • Each other catalog model has a maximum of 15 seconds, except where listed otherwise.

A small validation helper

Fetch the list once at startup, index by id, and validate duration, resolution and aspect_ratio before every submit. A catch at your edge gives a better error to your user.

Cache for minutes, not days: the catalog changes when Sume adds a model.

Known differences

size, seed and provider.options are rejected on these models, so a client should not send them. every v1 model reports no supported_sizes, which is why resolution plus aspect_ratio is the way to set output shape.

  • Read pricing_skus for list data.
  • Use GET /v1/catalog for the public catalog.
  • Record the response date.

Sources

Related posts

More in Developers

All Developers posts

Written by Sume