V

Vast.ai

Active

Low-cost GPU cloud marketplace

vast.ai · Founded 2020 · San Francisco, CA · Verified: 2026-03-24
8
Overall
9
Ease of Use
10
Pricing
10
GPU Variety
3
Enterprise

GPU Pricing

GPU ModelVRAMSpot $/hrOn-demand $/hrAvailable
RTX 509032GB$0.09$0.3 In Stock
RTX 409024GB$0.11$0.2 In Stock
H200141GB$1.94$2.19 In Stock
B200192GB$2.38$3 In Stock
H100 SXM80GB$0.45$1.47 In Stock
RTX PRO 6000 WS96GB$0.09$0.57 In Stock
A100 SXM480GB$0.27$0.53 In Stock
RTX 309024GB$0.02$0.11 In Stock
H200 NVL141GB$0.47$2.23 In Stock
L40S48GB$0.4$0.54 In Stock
A100 PCIE80GB$0.13$0.29 In Stock
H100 NVL94GB$0.47$1.6 In Stock
RTX A600048GB$0.13$0.33 In Stock
RTX 306012GB$0.03$0.04 In Stock
RTX 6000Ada48GB$0.27$0.44 In Stock
RTX A500024GB$0.02$0.13 In Stock
RTX 5060 Ti16GB$0.04$0.05 In Stock
H100 PCIE80GB$0.34$2 In Stock
A4048GB$0.23$0.29 In Stock
Tesla V10032GB$0.07$0.09 In Stock
RTX 5070 Ti16GB$0.07$0.09 In Stock
RTX PRO 400016GB$0.07$0.15 In Stock
RTX 4060 Ti8GB$0.01$0.09 In Stock
RTX 508016GB$0.08$0.13 In Stock
RTX 4070S Ti16GB$0.11$0.11 In Stock
RTX 4080S16GB$0.09$0.12 In Stock
RTX 3080 Ti12GB$0.06$0.07 In Stock
RTX 30708GB$0.04$0.05 In Stock
RTX 2080 Ti11GB$0.04$0.05 In Stock
RTX 407012GB$0.05$0.07 In Stock
RTX 4070 Ti12GB$0.05$0.06 In Stock
RTX 5880Ada48GB$0.23$0.33 In Stock
RTX PRO 400020GB$0.18$0.18 In Stock
RTX 507012GB$0.04$0.07 In Stock
Q RTX 500016GB$0.24$0.24 In Stock
RTX 4070S12GB$0.03$0.07 In Stock
RTX 3090 Ti24GB$0.07$0.14 In Stock
RTX 3060 Ti8GB$0.03$0.05 In Stock
RTX 408016GB$0.13$0.15 In Stock
GTX 10808GB$0.03$0.04 In Stock
GTX 1660 S6GB$0.04$0.04 In Stock
RTX A20006GB$0.03$0.05 In Stock
RTX 2060S8GB$0.05$0.05 In Stock
RTX 40608GB$0.06$0.06 In Stock
Quadro P4000$0.05$0.05 In Stock
Titan Xp12GB$0.03$0.04 In Stock
RTX 20606GB$0.03$0.04 In Stock
GTX 10708GB$0.01$0.37 In Stock
Tesla P48GB$0.02$0.02 In Stock
Q RTX 800048GB$0.25$0.21 In Stock
RTX 308010GB$0.04$0.08 In Stock
H10 PCIE$2$2 In Stock
RTX 6000ADA$0.47$0.47 In Stock
RTX 5880ADA$0.33$0.33 In Stock
RTX PRO 4000$0.17$0.17 In Stock
8200$3.06$3.06 In Stock
RTX PRO 6000 WS$0.79$0.79 In Stock
Q RTX 8000$0.24$0.24 In Stock
RTX PRO 4000$0.17$0.17 In Stock
RTX PRO 4000$0.17 In Stock
GTX 1070 Ti$0.03$0.03 In Stock
A10$0.2$0.2 In Stock
A800 PCIE80GB$0.4$0.88 In Stock
GTX 10606GB$0.07$0.09 In Stock
GTX 1070 Ti$0.03$0.03 In Stock
GTX 1080 Ti11GB$0.03$0.04 In Stock
GTX 16606GB$0.04$0.08 In Stock
GTX 1660 S$0.04$0.04 In Stock
GTX 1660 Ti$0.01$0.4 In Stock
L424GB$0.32$0.34 In Stock
L4048GB$0.47$0.58 In Stock
Q RTX 4000$0.03$0.04 In Stock
Q RTX 6000$0.16$0.13 In Stock
RTX 3060 laptop$0.05$0.08 In Stock
RTX 3070 Ti8GB$0.04$0.08 In Stock
RTX 4070S Ti$0.04$0.1 In Stock
RTX 4090D$0.33$0.43 In Stock
RTX 4500Ada$0.16$0.3 In Stock
RTX 5000Ada$0.17$0.3 In Stock
RTX 50608GB$0.07$0.11 In Stock
RTX PRO 5000$0.27$0.64 In Stock
Tesla M40$0.03$0.03 In Stock
Tesla P10016GB$0.09$0.08 In Stock
Tesla P4024GB$0.07$0.11 In Stock
Tesla T416GB$0.05$0.15 In Stock
Titan V$0.05$0.07 In Stock
RTX 2000Ada$0.05$0.07 In Stock
RTX 2070S8GB$0.03$0.09 In Stock
RTX 2080S$0.03$0.74 In Stock
RTX PRO 4500$0.1$0.23 In Stock

Features

Api
Docker
Jupyter
Kubernetes
Multi Gpu
Persistent Storage
Reserved Instances
Soc2 Compliant
Spot Instances

Billing & Payment

Billing Granularity

Per-Second

Payment Methods

Credit-Card, Crypto

Vast.ai is a GPU marketplace that flips the traditional cloud model on its head. Instead of a handful of data centers owned by a single company, Vast.ai aggregates GPU capacity from independent hosts worldwide — think everything from hobbyist rigs to small data centers — and lets you rent it all through one unified interface. The result is an absurdly wide selection of hardware at prices that undercut nearly every managed cloud provider.

Why Vast.ai stands out

The sheer breadth of hardware is genuinely unmatched. At any given moment you can find cutting-edge B200s and H100 SXMs sitting alongside budget-friendly RTX 3060s and even GTX 1070s. If you need a specific GPU for a specific budget, Vast.ai almost certainly has it. The marketplace model also means pricing is continuously competitive — hosts compete for your business, which keeps rates among the lowest available anywhere. Per-second billing ensures you’re never paying for idle time.

The spot market is particularly compelling for batch workloads. Interruptible instances go for even less than on-demand, and if your job checkpoints well, you can run significant training runs at a fraction of what you’d pay elsewhere.

Pros

  • Largest GPU variety of any cloud provider — 50+ models spanning every tier
  • Industry-leading pricing competitiveness; consistently among the cheapest options available
  • Per-second billing with no minimums
  • Spot instances for additional savings on interruptible workloads
  • Docker and Jupyter support out of the box
  • Accepts crypto payments
  • Excellent API for programmatic access and automation

Cons

  • No SOC2 compliance — not suitable for regulated workloads or enterprise security requirements
  • No Kubernetes support — container orchestration requires DIY setup
  • No reserved instances, so you can’t lock in capacity or pricing long-term
  • Host quality varies; uptime and reliability depend on individual hosts, not a managed SLA
  • Not enterprise-ready — support and accountability are limited compared to AWS or CoreWeave

Getting started

  1. Visit Vast.ai and create an account
  2. Add a payment method (credit card or crypto accepted)
  3. Use the search interface to filter GPUs by type, VRAM, price, and reliability score
  4. Select an instance and choose on-demand or spot pricing
  5. Connect via SSH or open a Jupyter notebook directly from the dashboard

Best for: Individual researchers, ML engineers, and startups who need maximum GPU variety at minimum cost, and can tolerate variable host reliability in exchange for the lowest prices in the market.

See something wrong? Report a data issue · DM on X