Models / FLUX.1 dev
Generate photorealistic avatars with FLUX.1 dev
The high-detail image model in Likna's lineup — the same FLUX.1 dev weights whether Likna renders on fal.ai's cloud (fal-ai/flux-lora) or on its own GPU. Samples below are plain FLUX.1 dev output; identity conditioning is a separate feature, on its own page. See sample output, pricing, and how it compares below.
Try FLUX.1 dev — freeSpeed
30 steps
Cost
2 credits / image (3 boosted to cloud)
Best for
Detail and realism
Sample output
How it compares
FAQ
What is FLUX.1 dev used for?
It's both the cloud model behind fal.ai's fal-ai/flux-lora endpoint and the base for Likna's local FLUX inference — the same underlying model either way.
Does it support Face Lock?
Yes — Face Lock (PuLID) runs on the FLUX family, so you can layer a locked reference face on top of FLUX.1 dev's output. It has its own page.
How many steps does it sample at?
30 by default, versus Z-Image Turbo's 8–10 — the extra steps are what buy the higher identity precision.
How much does an image cost?
2 credits, rendered on Likna's own GPU by default. Boost sends the same render to fal.ai's cloud instead for 3 credits total — the same Boost mechanism stills already use to skip the local queue.