MiniMax H3 languages: which spoken languages does it support?
MiniMax says H3 has stable support for 11 dialogue languages and partial support for others. Sume's docs list no language set, so test yours first.

MiniMax H3 has stable support for 11 dialogue languages, according to MiniMax: Arabic, Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian and Spanish. Other languages work "to varying degrees". Sume's docs do not list a language set for minimax-h3, so treat MiniMax's list as a guide and test your language.
As of 2026-09-29, the language list is on MiniMax's open-source announcement, in a specification table for the model. Sume's side comes from Video generation, which describes minimax-h3-max audio as native stereo and names no languages.
Which languages does MiniMax list?
| Item | What MiniMax's page says |
|---|---|
| Stable dialogue languages | 11: Arabic, Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, Spanish |
| Other languages | "Additional languages are also supported to varying degrees" |
| Output audio | 32 kHz stereo |
Does Sume say which languages H3 speaks?
No. The Sume docs describe minimax-h3-max as returning native stereo audio and say both H3 ids accept audio references, but they do not name languages, and this post does not add any. The MiniMax page describes its model release; it is not a promise about the hosted route Sume calls.
How do I test my language cheaply?
Run one short clip at the lowest resolution and listen. On Sume, minimax-h3 takes 5 to 15 seconds at native 480p or 768p, so a 5-second clip at 480p is priced at about $0.3125 (list × 1.25, plus a 5.5% agent fee by default). Write the line you want spoken into the prompt. Repeat the test once for each language you plan to ship, and keep the spoken line short so the result is easy to judge.
curl -X POST "https://api.sume.com/v1/videos" \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: h3-language-test-001" \
-d '{
"model": "minimax-h3",
"prompt": "A woman at a cafe table says a short greeting to the camera, natural room sound",
"resolution": "480p",
"duration": 5
}'Can I match a specific voice instead?
Sume's docs say audio references are honored by minimax-h3 and minimax-h3-max. A reference clip is the documented way to steer sound; see MiniMax H3 voice reference. It does not change which languages the model supports.
Do Sume's limits match MiniMax's page?
Not on clip length. MiniMax's specification table gives 4–15 seconds for the open release, while Sume's docs list 5–15 seconds for minimax-h3. Send 5 seconds or more, and trim afterwards if you need a shorter clip. Read each limit on the page of the product you call.
What if my language is not on the list?
Test it the same way, and judge the result yourself. Run the short clip above with the line written in your language, listen for the words you asked for, and only then plan a batch. MiniMax says other languages are supported to varying degrees, without naming them. For a fixed script, generating the speech separately is another route; AI voiceover for short videos covers it.
Sources
Related posts
More in Models
- MiniMax H3 Max 1080p: a latent refinement from native 768p
MiniMax H3 Max on Sume offers 480p, 768p and 1080p. The 1080p tier is a latent refinement from native 768p. What that means for price and output checks.
- What is MiniMax H3 Max? The post-trained variant, explained
MiniMax H3 Max is a variant post-trained by fal.ai on MiniMax H3 for faster generation. Its resolutions, lengths and modes, and how Sume lists it.
- MiniMax H3 Max lip sync API: a still plus audio, 5 to 14.8 seconds
Sume runs MiniMax H3 Max lip sync at POST /v1/minimax/h3-max/lip-sync: send a still and Sume-hosted audio of 5 to 14.8 seconds. Body, resolutions and price.
- MiniMax H3 open weights on Hugging Face: what is in the release
MiniMax H3 has a model card on Hugging Face: two checkpoints, a 33B-parameter Transformer, a community license and a 4-GPU serving example. What to check first.
Written by Sume