FLUX.1 Prompts: The Next Generation of AI Image Generation
A complete FLUX.1 Dev prompt guide. See how the battle of FLUX vs Midjourney v6 is shaping up, and explore incredible Black Forest Labs FLUX examples.

Meet the New King of Open Weights Developed by Black Forest Labs (a team containing original creators of Stable Diffusion), FLUX.1 has hit the scene like a freight train. It is currently the most capable open-weight image model in existence, and crafting **FLUX.1 prompts** requires a totally different mindset.
The FLUX.1 Dev Prompt Guide: Speak Naturally FLUX uses an incredibly massive text encoder (T5-XXL). This means it actually *reads and understands* your sentences, grammar, and spatial relationships.
In our FLUX.1 Dev prompt guide, the golden rule is simple: Write paragraphs, not tags.
Forget the comma-separated Midjourney style. Write like you are describing a scene to a blind sketch artist: > *A photorealistic image of a messy wooden desk. On the left side of the desk, there is a half-empty mug of black coffee with steam rising. In the center, a glowing smartphone screen displays the text "HELLO WORLD" in bold green font. On the right, a sleeping orange tabby cat is curled up next to a stack of vintage comic books. The lighting is soft morning sunlight coming through a window behind the desk, casting long shadows.*
FLUX will place the coffee on the left, the phone in the middle, and the cat on the right. Midjourney would likely blend the cat into the coffee.