fal
A generative media platform for developers to access models and deploy custom workloads on serverless GPUs.
Price: Premium
Description
fal is a generative media platform for developers, bringing image, video, audio, and 3D models together behind APIs. Its model gallery includes production-ready models that can be called through a unified API and SDKs, without requiring users to set up or fine-tune models first.
Developers can also deploy custom or fine-tuned models on serverless GPUs, or use dedicated compute clusters for training and inference. The platform describes its pricing as pay-per-use: Model APIs charge by model-specific units, while Compute uses hourly GPU pricing. fal also offers enterprise infrastructure features such as private deployments and single sign-on.
How to Use
1.Create an account and browse the model gallery for an image, video, audio, or other generative media model.
2.Review the selected model’s documentation and pricing details, including its billing unit and any settings that affect cost.
3.Call the model through fal’s API or SDK, passing the required inputs for your task.
4.For custom workloads, deploy a model with fal Serverless or choose dedicated Compute for training, fine-tuning, or inference.
Use Cases
Generate images or edit existing images through a model APICreate videos from text, images, or referencesBuild applications that use hosted generative media modelsDeploy custom or fine-tuned models for inferenceRun model training workloads on dedicated GPU clusters
Pros & Cons
Pros
- Access a broad gallery of image, video, audio, and 3D models through APIs
- Offers serverless GPU deployments without requiring users to configure GPUs or autoscaling
- Supports custom and fine-tuned models, including bring-your-own weights
- Provides dedicated GPU compute for training and inference workloads
- Pay-per-use pricing varies by model or GPU usage
Pricing
B300 Compute: $12.99/hour list price, as low as $5.99/hour - GPU with 288GB VRAM
GB200 Compute: $9.99/hour list price, as low as $5.89/hour - GPU with 192GB VRAM
B200 Compute: $7.99/hour list price, as low as $5.49/hour - GPU with 192GB VRAM
H200 Compute: $6.00/hour list price, as low as $2.99/hour - GPU with 141GB VRAM
H100 Compute: $4.50/hour list price, as low as $2.49/hour - GPU with 80GB VRAM
RTX PRO 6000 Compute: $4.00/hour list price, as low as $1.99/hour - GPU with 96GB VRAM
H3 Max video models: $0.05/second - Reference-to-video, image-to-video, and text-to-video endpoints
H3 Max Turbo video models: $0.025/second - Image-to-video and text-to-video endpoints
Kling Video v2.5 Turbo: $0.07/second - Image-to-video endpoint
Kling Video v3 Pro: $0.14/second - Image-to-video and text-to-video endpoints
Seedance 2 video models: $0.014/1000 tokens - Reference-to-video and image-to-video endpoints
Nano Banana 2: $0.08/image - Standard and edit endpoints
Flux 3 Image: $0.024/megapixel - Image editing endpoint
GPT Image 2 API: $1/unit - Standard and edit endpoints
Nano Banana Pro: $0.15/image - Standard and edit endpoints
Nano Banana: $0.0398/image - Standard and edit endpoints
Z Image Turbo: $0.005/megapixel - Turbo endpoint
FAQs
Related Tools
Open AI infrastructure for generating images, video, speech, music, and text, and building agents and apps.