Paperspace by DigitalOcean
Best for ML Notebooks, AI Model Development, Research, Computer Vision
from$0.08/ hr★ 8.6/10
Vast.ai
Available NowBest for Budget GPU Compute, Image Generation, Fine-Tuning, Batch Processing
About Vast.ai
Vast.ai is the world’s largest peer-to-peer GPU marketplace, founded in 2018 in San Francisco. Rather than owning data centers, Vast.ai aggregates spare GPU capacity from data centers and individuals globally, resulting in prices often 5-10x cheaper than traditional cloud providers.
How It Works
GPU Hosts list their GPUs and earn rental income. GPU Renters filter by DLPerf benchmark score, reliability rating, and location to find the best-fit machine at the lowest price.
Pricing Examples (2025)
- RTX 3090 24GB: from $0.09-0.25/hr
- RTX 4090 24GB: from $0.25-0.55/hr
- A100 80GB PCIe: from $0.99-1.99/hr
- H100 PCIe 80GB: from $2.49-3.50/hr
Pros & Cons
Pros
- Cheapest GPU cloud available
- Massive diversity of GPU types from global hosts
- Filter by reliability score, DLPerf benchmark, and location
- No minimum spend with pay-per-second billing
- Bid-based spot pricing for even lower costs
Cons
- Hardware reliability varies by host
- Not suitable for mission-critical production workloads
- No enterprise SLAs or compliance certifications
- Support is community-driven
Ideal Use Cases
AI InferenceFine-TuningImage GenerationResearch
| GPU Models | RTX 4090, RTX 4080, A100 80GB, H100 PCIe, A6000, RTX 3090, RTX 3080 Ti |
| GPU Types | A100, H100, RTX 3090, RTX 4090 |
| Headquarters | San Francisco, CA, USA |
| Founded | 2018 |
| Availability | Available Now |
| Website | vast.ai ↗ |
$0.09/ hour (starting)—$3.99/ hr (max)
💡 Pricing note: Rates shown are indicative. Final pricing depends on GPU model, reservation type (spot vs. on-demand), contract length, and region. Get an exact quote →
Global (100+ countries
decentralized peer-to-peer network)
Starting from
$0.09/hr
Up to $3.99/hr
Frequently Asked Questions
Vast.ai GPU cloud pricing starts from $0.09/hr depending on GPU type, reservation model (on-demand vs. spot vs. reserved), and region. Use the quote form to get exact pricing for your specific workload.
Vast.ai offers RTX 4090, RTX 4080, A100 80GB, H100 PCIe, A6000, RTX 3090, RTX 3080 Ti GPU instances. Availability varies by region and configuration. Contact the provider through ComputeStacker for current availability.
Vast.ai operates data centers in Asia Pacific, EU West, United States. Choosing a region close to your users minimises latency and can help with data residency compliance requirements.
Use the "Get a Quote" button on this page to submit your GPU requirements. ComputeStacker will forward your request to Vast.ai and other matching providers. You'll receive proposals within 24 hours — no commitment required.
Vast.ai offers high-performance GPU infrastructure suitable for large language model training and fine-tuning workloads. For large-scale distributed training, check the Specs tab for NVLink and InfiniBand interconnect availability.
