# of steps2
Inference time0.214s
Real-time
generative media
inference
Build the next generation of creativity with fal. Lightning fast inference.
Realtime Models
Sam 3SAM 3 is a unified foundation model for promptable segmentation in images and videos. It can detect, segment, and track objects using text or visual prompts such as points, boxes, and masks.
image-to-image
segmentation
rle
real-time
Sam 3SAM 3 is a unified foundation model for promptable segmentation in images and videos. It can detect, segment, and track objects using text or visual prompts such as points, boxes, and masks. vision
embeddings
mask
real-time
Sam 3SAM 3 is a unified foundation model for promptable segmentation in images and videos. It can detect, segment, and track objects using text or visual prompts such as points, boxes, and masks. video-to-video
segmentation
mask
real-time
rle
Sam 3SAM 3 is a unified foundation model for promptable segmentation in images and videos. It can detect, segment, and track objects using text or visual prompts such as points, boxes, and masks. video-to-video
segmentation
mask
real-time
Segment Anything Model 3SAM 3 is a unified foundation model for promptable segmentation in images and videos. It can detect, segment, and track objects using text or visual prompts such as points, boxes, and masks. image-to-image
segmentation
mask
real-time
Segment Anything Model 2SAM 2 is a model for segmenting images and videos in real-time.image-to-image
segmentation
mask
real-time
Segment Anything Model 2SAM 2 is a model for segmenting images and videos in real-time.video-to-video
segmentation
mask
real-time
MuseTalkMuseTalk is a real-time high quality audio-driven lip-syncing model. Use MuseTalk to animate a face with your own audio.image-to-video
animation
lip sync
real-time
Stable Diffusion XL LightningRun SDXL at the speed of lighttext-to-image
diffusion
lightning
real-time
Hyper SDXLHyper-charge SDXL's performance and creativity.text-to-image
diffusion
real-time
Latent Consistency Models (v1.5/XL)Run SDXL at the speed of lightimage-to-image
lcm
diffusion
turbo
real-time
editing
Latent Consistency Models (v1.5/XL)Run SDXL at the speed of lighttext-to-image
lcm
diffusion
turbo
real-time
Latent Consistency Models (v1.5/XL)Run SDXL at the speed of lightimage-to-image
lcm
diffusion
turbo
real-time
editing
Latent Consistency (SDXL & SDv1.5)Produce high-quality images with minimal inference steps.text-to-image
diffusion
lcm
real-time
Optimized Latent Consistency (SDv1.5)Produce high-quality images with minimal inference steps. Optimized for 512x512 input image size.image-to-image
diffusion
lcm
real-time
SDXL Realtime
This fast inference capability opens up new possibilities for application types that were previously not feasible, such as real-time creativity tools and using the camera as a real-time model input.
