Flux with LoRA and ControlNet — the model for structural control.
Flux General is the most controllable endpoint here. On top of LoRA support it accepts ControlNet conditioning, which means you can constrain the output to a structure you supply — a pose, a depth map, an edge outline — instead of describing that structure in words and hoping.
This is what you want when composition is non-negotiable: matching an existing layout, holding a figure in a specific pose across a series, or generating variations that all sit in the same architectural space. Prompting cannot give you that reliably; conditioning can.
It expects a conditioning input to be worth the price. Used as a plain text-to-image model it is Flux Dev at the same cost with more knobs.
| Price | 210 credits — about 19c |
|---|---|
| Type | Image |
| Provider | fal |
| Aspect ratios (13) | 1:1, 4:3, 16:9, 3:2, 19.5:9, 20:9, 2:1, 3:4, 9:16, 2:3, 9:19.5, 9:20, 1:2 |
| Resolutions | 1k, 2k |
| LoRA | Yes — Hugging Face or CivitAI adapter URLs |
Credit prices are the live studio rates. Cash equivalents are approximate at starter-pack rates and improve on the larger packs — see pricing.
Every AI model on comfyarts — image models, prices and measured capabilities side by side.