Models
Five models for image, video, speech and 3D. Each one is billed per output, and a failed run is refunded in full.
FLUX.1 [schnell]
Fast text-to-image in 1–4 steps. Good default for drafts and thumbnails.
$0.0042
per megapixel · Black Forest Labs
Z-Image Turbo
Photoreal text-to-image with strong prompt following and legible text.
$0.0063
per megapixel · Tongyi-MAI
Wan 2.2 5B
Text-to-video clips, about 5 seconds at 720p.
$0.2100
per video · Wan-AI
fal retires this model on 2027-01-04. Swap the endpoint in src/lib/models.ts before then.
MiniMax Speech 02 Turbo
Natural text-to-speech. Billed by characters.
$0.0840
per 1,000 characters · MiniMax
TRELLIS
Image to textured 3D asset (GLB).
$0.0280
per asset · Microsoft