How does pricing work on Vast.ai?

When an instance is stopped on Vast.ai, compute charges cease, but attached storage continues billing at $0.02 - $0.05/GB/month until the disk is deleted.

How much can I save on Vast.ai?

Vast.ai spot instances typically provide 20% to 60% discounts compared to standard on-demand pricing.

Vast.ai Cost & Storage Simulator

Verified Formula

Simulate real compute, idle storage, and monthly bill estimates

Live Rates Table Compare Rivals Deploy on Vast.ai

Interactive On-Site Simulation

Vast.ai Monthly Cost & Stopped Storage Estimator

Simulate your real-world monthly bill including active GPU compute and stopped persistent volume storage overhead.

Full Multi-Cloud Calculator

Workload Presets:

1. Select GPU Hardware Tier1x NVIDIA GeForce RTX 4090 (24GB VRAM)

2. Pricing Tier

3. Active Daily Compute Hours6 hrs / day (183 hrs / mo)

1 hr (Testing)8 hrs (Full workday)16 hrs (Heavy training)24 hrs (Continuous serving)

4. Attached Persistent Volume Disk200 GB NVMe SSD

50 GB (Base OS)250 GB (ComfyUI + Models)500 GB (Flux + SDXL cache)1,000 GB (1TB 70B Weights)

Developer Gotcha: Stopped Storage Billing

When you click Stop Instance on Vast.ai, your GPU compute charge stops, but the 200GB disk volume continues billing 24/7 ($20.00/mo) until you terminate the volume. Always backup checkpoints to HuggingFace or S3 before terminating.

Estimated Total Monthly BillSpot Rate

$56.60 / month

Hourly run-rate: $0.20/hr compute + $0.0274/hr disk

Active GPU (183 hrs):$36.60

Persistent Disk (200GB 24/7):$20.00

You save ~$189/mo (77%) vs AWS EC2

Equivalent cloud capacity on AWS would cost approx. $245/mo for 6h daily runtime.

Compare NVIDIA GeForce RTX 4090 Rates Across 15 Clouds

Full Multi-Cloud Calc VS Battles Matrix

Estimated Monthly Cost for this Workload:

Vast.ai:$56.60/mo

RunPod Equivalent:$100.52/mo

AWS Cloud Equivalent:$245.43/mo

The Stopped Storage Fee Gotcha on Vast.ai

When you stop an instance on Vast.ai, your hourly GPU compute charge drops to $0.00/hr immediately. However, any persistent disk volume (where your ComfyUI models, LoRAs, or Conda environments live) remains reserved on physical NVMe drives and continues accruing storage billing 24 hours a day, 7 days a week ($0.02 - $0.05/GB/month).

100 GB Idle Volume~$10.00 / monthContinuous stopped retention

500 GB Full Weight Volume~$50.00 / monthEquivalent to ~100 hrs RTX 4090 spot

Cloudflare R2 Backup$0.015 / GB/mo (85% Cheaper)Zero egress fees for weights sync

100-Hour Monthly AI Spend Benchmark vs Other Clouds

Scenario: 100 hrs RTX 4090 + 200 GB persistent storage

Vast.ai

~$40 - $70

Includes spot discount + storage fee

RunPod

~$54 - $74

$0.34/hr spot + $0.10/GB/mo network volume

TensorDock

~$39 - $59

$0.19/hr spot base + $0.08/GB/mo storage

AWS EC2 (G5)

~$120 - $160

A10G equivalent + EBS gp3 disk + egress fees