gallery View Gallery

Vast AI

Be the first to Reviews

Have Questions?

Get Free Expert Consultation

Let us know your business needs clearly

Verified & Trusted
Vendors

Safe & Secure
Payment

Personalized
Experts Support

Price On Request

Features

See all

Vast AI Features

Multiple instance types Supports on-demand, reserved, and interruptible (spot) GPU instances.
Interruptible instances Use cheaper compute that can be preempted when outbid.
Persistent storage billing Even when a GPU instance is stopped, storage cost continues until deletion.
Bandwidth based charges Data transfer (upload/download) is billed per byte based on host bandwidth rates.
Custom Docker images Build your image (or use base ones), push to registry, and launch via template.
Launch modes Choose between SSH, Jupyter, or custom entrypoint for how you connect to an instance.
Global GPU fleet search filters Filter offerings by GPU model, RAM, CPU cores, bandwidth, and more.
Serverless auto scaling Automatically scale GPU workers based on load via the serverless system.
PyWorker integration A Python web server (PyWorker) runs on each worker to serve inference and report metrics.
Performance benchmarking Workers run a performance test on startup to measure GPU throughput (e.g., tokens/s).
Logs and debugging Detailed logs for endpoints and worker groups, with debug/trace/info levels.
Teams and shared billing Team feature for shared resources, access control, and consolidated billing.

Brand Name	Vast AI
Information	Vast.ai is the market leader for low cost GPU rentals. The service connects data centers and professionals running the Vast hosting software with users who can quickly find the best deals for compute according to their specific requirements.
Founded Year	2018
Director/Founders	Jake Cannell
Company Size	1-100 Employees