# Vast.ai vs RunPod

Canonical: https://inetgeek.com/compare/vast-ai-vs-runpod/

Every value below is read from Vast.ai's and RunPod's own documentation. See https://inetgeek.com/methodology/ for how.

## At a glance

| Criterion | Vast.ai | RunPod |
| --- | --- | --- |
| H100 SXM, per GPU-hour | Median $2.16 per hour for H100 SXM 80GB, with a floor of $1.33 from the cheapest host. Vast is a marketplace — the floor is one host's offer at one moment and the median is what the market is actually charging, so both move without anyone announcing anything. This figure is a snapshot, not a rate card. | $3.49 per hour for H100 SXM 80GB, single card. The PCIe variant is $2.89 and H100 NVL is $3.19. |
| Cheapest GPU, per hour | Floor and median vary by card and by the hour; H100 PCIe runs from $1.87 with a median of $2.67, and H200 from $1.97 with a median of $4.24. The H200 floor now sits below the H100 PCIe median, which is the kind of inversion a marketplace produces and a rate card cannot. | $0.27 per hour for an RTX A5000 with 24GB. RTX 3090 is $0.50, RTX 4090 $0.74, L40S $1.09. |
| Largest GPU offered | 192GB per GPU on B200, from $4.38 an hour with a median of $7.50. | 288GB per GPU on B300; B200 at 180GB is $6.79 an hour and H200 at 141GB is $4.59. |

## Where they differ

### H100 SXM, per GPU-hour

- Vast.ai: Median $2.16 per hour for H100 SXM 80GB, with a floor of $1.33 from the cheapest host. Vast is a marketplace — the floor is one host's offer at one moment and the median is what the market is actually charging, so both move without anyone announcing anything. This figure is a snapshot, not a rate card. ([source](https://vast.ai/pricing))
- RunPod: $3.49 per hour for H100 SXM 80GB, single card. The PCIe variant is $2.89 and H100 NVL is $3.19. ([source](https://www.runpod.io/pricing))

### Cheapest GPU, per hour

- Vast.ai: Floor and median vary by card and by the hour; H100 PCIe runs from $1.87 with a median of $2.67, and H200 from $1.97 with a median of $4.24. The H200 floor now sits below the H100 PCIe median, which is the kind of inversion a marketplace produces and a rate card cannot. ([source](https://vast.ai/pricing))
- RunPod: $0.27 per hour for an RTX A5000 with 24GB. RTX 3090 is $0.50, RTX 4090 $0.74, L40S $1.09. ([source](https://www.runpod.io/pricing))

### Largest GPU offered

- Vast.ai: 192GB per GPU on B200, from $4.38 an hour with a median of $7.50. ([source](https://vast.ai/pricing))
- RunPod: 288GB per GPU on B300; B200 at 180GB is $6.79 an hour and H200 at 141GB is $4.59. ([source](https://www.runpod.io/pricing))

## Which should you choose?

Pick Vast.ai if Price-sensitive training and batch inference that can tolerate a marketplace — the median H100 rate is well under half what the managed clouds charge, and interruptible capacity goes lower still.

Pick RunPod if Short experiments and bursty inference — per-second billing and a $0.27 entry card mean an idle hour costs nothing and a small job is genuinely small.

Consider something else: Vast.ai — There is no single price to plan against: rates move with supply, the host is not Vast itself, and an interruptible instance can be reclaimed mid-run.

Consider something else: RunPod — The catalogue mixes community and secure capacity, so the cheapest quoted rate is not always the tier a production workload should sit on.

## Documented by only one

- Pricing model: Vast.ai: A marketplace rather than a cloud: prices are set by supply and demand across more than 40 data centres, and the same card can be rented on demand, interruptibly, or reserved. Rates move continuously, which is why every figure here carries a 7-day re-verification window rather than the category's 30.
- Storage price: RunPod: Persistent storage from $0.05 per GB per month, in standard and high-performance tiers.
- Billing granularity: RunPod: Per second, with per-hour rates shown for comparison. Clusters can be started with no commitment and scaled to 64 GPUs.

## Questions this comparison answers

**Should I choose Vast.ai or RunPod?**

Pick Vast.ai if Price-sensitive training and batch inference that can tolerate a marketplace — the median H100 rate is well under half what the managed clouds charge, and interruptible capacity goes lower still.
Pick RunPod if Short experiments and bursty inference — per-second billing and a $0.27 entry card mean an idle hour costs nothing and a small job is genuinely small.