Gemini Omni and Veo in Korean: only English is fully supported
Google's Omni and Veo pages say English is fully supported; other languages aren't evaluated. For Korean prompts, describe in English and quote on-screen text.

Can you prompt Gemini Omni Flash or Veo in Korean? You can try, but Google does not promise it. Both of its video pages carry the same sentence: English is fully supported, but other languages have not been evaluated, so they may work but results can vary. For a Korean-language ad, the safer pattern is an English scene description with any on-screen words given exactly, in quotes.
The line is on Google's Gemini API: Generate and edit videos with Gemini Omni Flash and Gemini API: Generate videos with Veo, both read 2026-10-02. Sume's Video Router lists gemini-omni-flash-1.1 and does not state a prompt-language rule, so Sume adds no promise beyond Google's.
What does Google's page actually say?
The wording sits in the technical details of the Omni guide and in the limitations of the Veo guide: "English (EN) is fully supported, but other languages have not been evaluated, so they may work but results can vary." That is a statement about evaluation, not a statement that other languages fail.
Nothing on either page lists which languages were tried, and neither gives a quality figure. Treat Korean as untested by Google rather than as broken.
| Model | Language line on Google's page | Where |
|---|---|---|
| Gemini Omni Flash | English fully supported; others not evaluated, results can vary | Technical details |
| Veo 3.1 and family | Same sentence | Limitations |
How should I write the prompt?
Google's Omni guide has a section on text in videos: you can prompt for text and it will render it, and if text appears naturally in the scene, even in the background, it helps to define what it should say. Its examples put the exact words in quotes, such as a street sign that says "This is an AI generation by Omni".
So describe the shot in English, and put the exact Korean wording in quotes where it must appear on screen. Check the rendered characters before you publish; the page makes no promise about non-English text rendering.
What about spoken dialogue?
Google's Veo guide says dialogue goes in quotes inside the prompt, and the Omni guide suggests describing the audio you want, for example calm background music. Neither page says the spoken language will match the language of the prompt. If a clip needs reliable Korean speech, plan to test it on a short, low-resolution draft before spending on a long one.
What should I do on Sume?
The Video Router page lists 3 to 10 seconds at 360p, 720p, 1080p and 4K for Omni, so start a test at the lowest resolution. Send the same prompt shape you would send to Google, and check the output before batch work.
Sources
Related posts
More in Models
- Gemini Omni can't use a YouTube link as its source video
Google lists YouTube videos as an unsupported media source for Gemini Omni Flash. On Sume, video_url is a URI field; use a direct link to the clip file.
- GPT Image 2.5 mask_url edit: change one region, keep the rest
GPT Image 2.5 on Sume accepts a public mask_url alongside input_references. Build a mask with Pillow, host it, and edit only one region of a photo.
- GPT Image 2.5 character drift: what OpenAI says and what to send
OpenAI's image guide says GPT Image may struggle to keep a recurring character the same across generations. What to send on Sume to reduce drift.
- gpt-image-2.5 smallest custom size: the 655,360-pixel floor
gpt-image-2.5 on Sume rejects images under 655,360 pixels. Why 512x512 and 768x768 fail, which small sizes pass, and the other edge rules to check first.
Written by Sume