FLUX.2 flex text rendering: prompt tips and the Sume model id

FLUX.2 [flex] is the variant Black Forest Labs says "excels at rendering text and fine details", with developer-controlled parameters. On Sume you call it as black-forest-labs/flux.2-flex through POST /v1/images; Sume's docs do not expose the steps and guidance settings BFL describes, so text quality here comes from the prompt.

Vendor claims are from Black Forest Labs' launch post and typography guide; Sume's from the Image API docs, all read 2026-09-29.

How should I write text into a FLUX.2 prompt?

BFL's guide gives four habits. They are prompt-writing advice from the vendor, and results vary by image.

• Put the exact text in quotation marks, for example: The text 'OPEN' appears in red neon letters above the door.
• Say where the text sits relative to other elements.
• Describe the style: serif, industrial lettering, handwritten script.
• Give a size cue such as large headline or small body copy, and a hex code for a brand color: the logo text 'ACME' in color #FF5733.

Can I set steps and guidance on Sume?

Black Forest Labs documents [flex] steps up to 50 and guidance 1.5 to 10 on its own API. Sume's image docs list allowed_passthrough_parameters as empty for every endpoint in v1, so provider.options must be omitted or empty and those settings are not available through Sume.

What does a flex text request look like?

curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "black-forest-labs/flux.2-flex",
    "prompt": "A cafe chalkboard. The text SPRING MENU in white handwritten script at the top, large.",
    "aspect_ratio": "3:4"
  }'

What does one flex image cost?