FLUX 3 Image

Black Forest Labs’ FLUX 3 Image: new images at 1K, 2K or 4K in 15 shapes, legible text in the picture, and edits of an image you send.

Price

1K $0.04, 2K $0.05, 4K $0.09 an image. An edit costs the same as a new image, and the image you send is included.

Make an image

import os
from openai import OpenAI
client = OpenAI(base_url="https://api.zurelay.com/v1", api_key=os.environ["ZURELAY_API_KEY"])
result = client.images.generate(
    model="flux-3-image",
    prompt="A poster for a night market: lanterns over food stalls in the rain, the words 'NIGHT MARKET' in bold letters across the top and 'Fridays 6 pm to midnight' below",
    size="2K",
    extra_body={
      "aspect_ratio": "2:3"
    },
)
print(result.data[0].url)

Parameters

FLUX 3’s sizes have the same area at every shape: about one megapixel at 1K (1024 × 1024 square), four at 2K (2048 × 2048, or 2736 × 1536 for 16:9) and sixteen at 4K (5472 × 3072 for 16:9). A 1K image in a shape other than square can come back at the 2K size, at the 1K price. An image smaller than asked, or in another shape, is retried, never delivered.

Edit an image

import os
import requests
response = requests.post(
    "https://api.zurelay.com/v1/images/generations",
    headers={"Authorization": f"Bearer {os.environ['ZURELAY_API_KEY']}"},
    json={
      "model": "flux-3-image",
      "prompt": "Make the neon sign glow red instead of blue. Keep everything else exactly the same.",
      "images": [
        "https://example.com/storefront.png"
      ]
    },
)
response.raise_for_status()
print(response.json())

Prompting

• Describe the subject, the setting, the light and the medium in full sentences.
• Put words that must appear in quotes and say where they go.
• There’s no negative prompt: describe what should be there (“an empty street”) rather than what shouldn’t (“no people”).
• For edits, name what changes and what stays: “Change the sweater to deep blue. Keep everything else the same.”

Good to know

• Images take 30 seconds to 2 minutes at 1K and 2K, and 3 to 4 minutes at 4K.
• The model can look things up on the web while it draws, for prompts about recent events.
• Prompts the model declines return image_not_generated, free of charge.
• Links last 24 hours; the Library in your dashboard keeps images returned as links for 30 days.

Model details: FLUX 3 Image. More on sending images: Image editing and reference images.