FLUX.2 [flex] - Starrise AI

FLUX.2 [flex] is a highly configurable text-to-image model from Black Forest Labs, available through Starrise AI. It offers fine-grained control over sampling steps and guidance scale, ideal for design professionals. The API uses synchronous calls — it directly returns Base64-encoded image data.

Key Capabilities

Text-to-image — Generate images from text descriptions
Step control — Adjust sampling steps (1–50) to flexibly balance speed and quality
Guidance scale — Fine-tune prompt adherence (1.5–10)
High resolution — supports up to 4MP output resolution
Reproducible — Use seed for deterministic generation

Quick Example

curl https://ai.alad.com/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "flux-2-flex",
    "prompt": "A cute baby polar bear",
    "size": "1024x1024",
    "inference_steps": 30,
    "guidance_scale": 7.5
  }'

Parameters

Parameter	Type	Required	Description
`model`	string	Yes	Fixed value: `flux-2-flex`
`prompt`	string	Yes	Text description for the image
`n`	integer	No	Number of images to generate (1–4), default `1`
`size`	string	No	Output resolution, e.g. `1024x1024`, supports up to 4MP
`inference_steps`	integer	No	Sampling steps (1–50). Low steps (6–20) for rapid prototyping, high (40–50) for max fidelity
`guidance_scale`	number	No	Guidance scale (1.5–10). Higher values produce closer adherence to the prompt
`output_format`	string	No	`jpeg` or `png`
`seed`	integer	No	Random seed for reproducibility

Response Fields

Field	Description
`data[].b64_json`	Base64-encoded image data
`created`	Creation timestamp

API Reference

View the interactive API Playground for FLUX.2 [flex].

FLUX.2 [pro]FLUX.1 Kontext [pro]

⌘I