Open-weights video news does not change your Sume bill: LTX, H3, Wan
LTX-2.5 and MiniMax H3 are reported open weights, but Sume still bills hosted per-second rows. Which ids to call and what each costs.
Open weights were reported for LTX-2.5 and for MiniMax H3 (reported by Magic Hour, read 2026-10-07), and Wan 3.0 and Seedance 2.5 are reported at the top of the leaderboard (reported by Pinggy, read 2026-10-07). On Sume, the model ids that exist are hosted. Prices are Sume billable rates (provider list x 1.25, rounded up to the cent per job), read 2026-10-07.
What to call
| Name in the news | On Sume | Cheapest row |
|---|---|---|
| LTX-2.5 | no id | use Wan 3.0 or H3 ($0.32 for 5 s at 480p) |
| HunyuanVideo | no id | same |
| MiniMax H3 | minimax-h3 | $0.0625 at 480p |
| Wan 3.0 | wan-3.0 | $0.0625 at 480p |
| Seedance 2.5 | seedance-2.5 | per 1000 tokens |
Open weights lower the cost of self-hosting; they do not change a hosted per-second rate. The break-even post linked below gives the formula for deciding.
Sources
Related posts
More in Models
- Pick a video model for a Reel by length: 3, 10, 15 or 30 seconds
The length you need narrows the Sume models fast: 2 s is Wan only, 10 s allows Omni, 30 s allows Seedance 2.5 or Wan. Table of the shortlist.
- Portrait 4K AI video on Sume: Omni Flash 9:16 at $0.375 a second
LTX-2.5 is reported to do portrait up to 4K. On Sume the 4K portrait option is Gemini Omni Flash 1.1 at $0.375 a second: 3 s $1.13, 10 s $3.75.
- Qwen Image 2.1 sizes (2752x1536, 2400x1792) vs Sume Qwen ratios
The Qwen-Image-2.1 card lists 2048x2048, 2400x1792 and 2752x1536. Sume qwen-image takes 13 aspect_ratio values instead of pixels. How they line up.
- Qwen Image 2.1 Pro API: the model card lists no Pro variant
Searching for a Qwen Image 2.1 Pro API? The 2.1 model card names no Pro variant and no hosted endpoint. Here is what Sume lists for Qwen today.
Written by Sume