"A [product] on [surface], [light]. A small label that reads exactly "[TEXT]". No other text."
In our test it set the label exactly as written, spacing included.
ByteDance's Seedream 5.0 Pro: the full model of the Seedream 5.0 family, beside Flash and Lite. FlowPix makes a 1K picture unless you pick 2K, and it goes no higher. It takes up to 10 reference images and edits pictures you give it. FlowPix uses it when you pick or name it: the default stays GPT Image 2.5.
Our two test pictures at 1K came back in 40 s and 41 s. The label read exactly MORNING ROAST · 250 g, the space included, where both GPT Image 2.6 versions dropped it. The fisherman portrait has the rain drops, mist and shallow depth of field the prompt asked for. Two prompts are a small test: we have not tried 2K, references or edits.


Example prompts stay in English, which is what the models handle best.
"A [product] on [surface], [light]. A small label that reads exactly "[TEXT]". No other text."
In our test it set the label exactly as written, spacing included.
"[Subject and scene], [light], [lens or depth of field]. — then pick 2K."
1K is the default; pick 2K for a larger picture.
Yes, up to 10. It also edits a picture you give it. The credits a picture costs are shown when you choose to generate.
2K. FlowPix offers 1K, the default, and 2K, and it goes no higher. For a larger file, upscale the picture afterwards.
Flash is the lowest-priced Seedream 5.0 and makes 1K pictures. Lite is the artistic one. Pro is the full model, at 1K or 2K.
When your request plays to Seedream 5.0 Pro's strengths. Name it in your message or pick it in the composer; FlowPix never switches models on its own, and uses its default image model when you choose none.
Yes. The model picker in any chat lets you pin per-turn or per-session. Set as default in /settings.