FLUX.2 native-language prompts: Japanese or Korean vs English
BFL says native-language FLUX.2 prompts often look more culturally authentic. Send both versions to a Sume FLUX.2 id with n and compare the results.
Yes, try the native-language prompt. BFL's guide says FLUX.2 understands multiple languages and that prompting in the native language of the content you are creating often produces more culturally authentic results than a translated English prompt. Sume's image API takes the prompt as a plain string, so you send the Japanese or Korean text unchanged and compare it against an English version.
The vendor claim is from BFL's Multi-Language Prompting page; request rules from Sume's Image API docs, read 2026-10-01. The word "often" is BFL's; test on your own subject.
What does BFL actually claim?
Local markets, architecture and atmosphere are, in BFL's words, rendered with greater accuracy than with translated English prompts. The page shows examples as a prompt carousel; it gives no measured numbers, so do not treat it as a benchmark.
How do I run the comparison on Sume?
| Step | Field or action |
|---|---|
| Choose a model | A FLUX.2 catalog id such as black-forest-labs/flux.2-pro |
| Two requests | Same model, one native-language prompt, one English prompt |
| Variants | n takes up to 10 images per request; per-model ceilings are lower, so read the catalog |
| Same settings | Keep aspect_ratio identical so only the language differs |
How do I confirm the model is available?
You can explore available models and pricing with GET /v1/images/models, which lists what the catalog serves today. Use an id from that list rather than one copied from a vendor page.
What can go wrong?
A parameter the model does not list is rejected with 400 unsupported_parameter rather than silently dropped, so a field copied from another provider's docs may fail. Text drawn inside the image is a separate question; see image text rendering models on Sume.
Sources
Related posts
More in Models
- FLUX 3 Action is a robot-control model, not an image model
FLUX 3 Action is an open-weight 7B world action model for robots. It predicts actions, not images, so it is no drop-in for the FLUX ids Sume serves.
- FLUX 3 Image bounding boxes vs Sume masked edit: what each takes
FLUX 3 Image takes a JSON list of boxes at the end of the prompt. Sume's image edit takes a mask image: mask_image_url with image_urls. Differences, no overlap.
- FLUX 3 Image 10 reference images vs Sume's 1-10 image_urls
FLUX 3 Image takes up to ten references named by prompt position. Sume's image edit takes one to ten public image_urls, with no FLUX 3 id in the docs.
- FLUX 3 x mimic: one backbone for video, audio and robot actions
BFL says FLUX 3 is one model trained jointly on images, video and audio, and FLUX-mimic adds robot actions. What that does and does not tell a media API user.
Written by Sume