Mingmou 1 and Hunyuan 1.5 on Sume: fixed lengths and low rates

Sume's Tencent rows Mingmou 1 and Hunyuan 1.5 bill $0.05775 per second at 720p, take 5 or 10 seconds (Hunyuan 5 only), no audio. What they can and cannot do.

4 min readSume
All posts

Mingmou 1 and Hunyuan 1.5 are Tencent-hosted text-to-video rows on Sume that bill $0.05775 per second at 720p, with fixed lengths: Mingmou takes 5 or 10 seconds in 16:9, and Hunyuan takes 5 seconds in 16:9 or 9:16. Neither produces audio.

They appear in a workspace only when GET /v1/videos/models returns their ids. All figures here come from Sume's pricing code, billed as list times 1.25.

What the pricing code says

In the Tencent VOD spec in the pricing package, both rows share one rate card across four resolutions and have audio set to none, so generate_audio does not apply. Durations are validated against an explicit list rather than a range, which means 7 seconds is rejected even though it lies between 5 and 10.

That is the most common mistake: sending a duration that other rows accept. Read supported_durations in the catalog and send only a value in it.

Mingmou 1 and Hunyuan 1.5 limits (Sume pricing code, checked 2026-10-10)
RowDurationsAspect ratiosAudio
mingmou-15 or 10 s16:9None
hunyuan-1.55 s16:9, 9:16None

Billed rates and clip costs

Both rows bill the same per-second rate at each resolution. The table gives the per-second rate and the cost of the allowed clips: 5 and 10 seconds for Mingmou, 5 seconds for Hunyuan.

A 10-second Mingmou clip at 720p therefore bills $0.5775, about the price of a 5-second Wan 3.0 clip at 720p ($0.625).

Billed per second and per clip on Sume (list x 1.25, checked 2026-10-10)
ResolutionPer second5 s clip10 s clip (Mingmou)
720p$0.05775$0.28875$0.5775
1080p$0.09625$0.48125$0.9625
2K$0.144375$0.721875$1.44375
4K$0.215625$1.078125$2.15625

Where they fit

They suit cheap text-only drafts and stock-style filler where silence is fine and a vertical or landscape clip of exactly 5 seconds is enough. A 4K 5-second clip bills $1.078125, which is below a 5-second Seedance 2.5 clip at 480p ($1.343385) on Sume.

They do not suit anything that needs references, an end frame, audio or a custom length. For those, use Wan 3.0, Seedance 2.5 or Gemini Omni Flash 1.1, whose limits are in the Video Generation docs.

  • Need 21:9, 4:3, 1:1 or 3:4: not on these two rows; WAND-Vega covers them.
  • Need audio: use a row whose catalog entry has generate_audio true.
  • Need 9:16 at 10 seconds: Mingmou is 16:9 only and Hunyuan is 5 s only, so use another row.

Practical advice

Because the lengths are fixed, plan shots in 5-second units. A 20-second sequence is four requests, each billed separately, and each should carry its own Idempotency-Key. If one of them fails with a 503 provider_capacity_exceeded, retry with the same key, as the Errors and rate limits page describes.

Cost per ten drafts

A quick way to see the appeal is to price a batch. Ten 5-second 720p drafts on either row bill $2.8875 in total. Ten 5-second 720p drafts on Wan 3.0 bill $6.25, and on Seedance 2.5 at 480p they bill $13.43385. At 720p the Tencent rows undercut Wan 3.0 and Seedance 2.5, if your workspace lists them.

The saving buys you less control: no audio, no references, no end frame, and one or two fixed lengths. Use them to test ideas, then move the winning prompt to a row that gives you the controls the final needs.

Sources

Related posts

More in Models

All Models posts

Written by Sume