Video license size lines: 100M users, $20M and $10M revenue
Hunyuan's line is 100 million monthly users, H3's is US$20 million revenue, LTX-2.5's is $10 million with affiliates counted. Wan 2.2 has none.
Four open video models set four different size lines. HunyuanVideo asks for a Tencent license above 100 million monthly active users. MiniMax H3 asks for written authorization above US$20 million in yearly revenue. LTX-2.5 asks for a paid license from entities with annual revenues of at least $10,000,000, counting subsidiaries and affiliates together. Wan 2.2 is Apache 2.0 and the license has no size line. These come from each vendor's own file, read 2026-10-08.
Side by side
A user-count line and a revenue line are not comparable by size. A pre-revenue app with 150 million monthly users trips Hunyuan's line and no other. A $15 million agency with a few thousand users trips LTX-2.5's line and H3's line only if revenue passes each number.
| Model | Metric | Line | What happens above it |
|---|---|---|---|
| HunyuanVideo | Monthly active users of all your products, on the version release date, in the preceding calendar month | More than 100 million | You must request a license; no rights until Tencent grants one |
| MiniMax H3 | Yearly revenue of commercial products or services | Over US$20 million | Separate written authorization from MiniMax |
| LTX-2.5 | Annual revenues of the Entity, affiliates included | At least $10,000,000 | A paid Commercial Use Agreement is required |
| Wan 2.2 | None | None | Apache 2.0 conditions only |
Details that change the answer
Hunyuan measures users on the release date of the version, with all products 'made available by or for Licensee'. LTX defines Entity to include subsidiaries, affiliates and companies under common Control, so a small studio inside a large group counts the group. H3's page also asks that commercial products show 'MiniMax H3' prominently in the user interface and says outputs may not be used to train other AI models.
LTX's August 2026 text adds a carve-out: a Commercial Entity may use LTX-2.x for a defined non-commercial purpose without the paid license, such as testing or evaluation in a non-production environment. Revenue-generating use or use with direct end-user impact is not covered by that carve-out.
Where a hosted call avoids the question
None of these size lines apply to a team that calls a hosted API, because the team is not licensing weights. The cost of that choice is a per-second price. On Sume, a ten-second 768p clip is $0.75 on minimax-h3 and $1.25 on wan-3.0 at 720p, using the list x 1.25 rule in Sume's video docs.
Sume does not list HunyuanVideo or LTX-2.5, so it cannot substitute for those two. It lists the H3 and Wan 3.0 families.
A quick screen
Ask three questions in order:
- Do we, with affiliates, earn at least $10 million a year? If yes, LTX needs a paid agreement.
- Do we earn over US$20 million from products that would use H3? If yes, write to MiniMax.
- Do our products reach 100 million monthly users? If yes, Hunyuan needs a Tencent license first.
Sources
Related posts
More in Models
- 1,000 GPT Image 2.5 images: $10 on low up to $270 on max, on Sume
Cost of 100, 500 and 1,000 GPT Image 2.5 images at each quality tier on Sume: low, medium, high, xhigh and max, with the arithmetic behind xhigh and max.
- 10,000 images on Sume: $100 on Soul up to $2,700 on ChatGPT Image 2
What 100, 500, 1,000 and 10,000 images cost on each of the 19 models in the Sume image catalog, with the per-image arithmetic shown.
- 16 references on GPT Image 2.5, 5 on Ideogram 4.5: read limits in code
GPT Image 2.5 accepts up to 16 input_references and Ideogram 4.5 accepts 5 in total. A short Python script reads every model's reference limit from the catalog.
- 20 s Seedance: only 2.5 passes 15 s, $11.56 vs two 2.0 jobs at $7.56
Seedance 2.0 stops at 15 s on Sume; 2.5 takes 4 to 30 s. At 720p 9:16, 20 s on 2.5 is $11.56, two 10 s 2.0 jobs $7.56, one 20 s Wan 3.0 job $2.50.
Written by Sume