Vast.ai GPU Rental Pricing 2026: Marketplace Rates

Quick Answer

Vast.ai GPU rental pricing 2026: H100 SXM marketplace floor $1.33/GPU-hr interruptible; H100 PCIe unverified $0.90-$1.60, datacenter-verified $1.50-$2.27; H200 verified $3.50-$5.50; B200 intermittent $6-$12; A100 80GB $0.79/GPU-hr. Per-second billing, host-level reliability. The cheapest published GPU rates in the category.

Last verified: Sep 16, 2026.

At a glance

• H100 PCIe unverified: $0.90-$1.60/GPU-hr (host-level reliability variance)
• H100 PCIe verified: $1.50-$2.27/GPU-hr (datacenter-vetted)
• H100 SXM verified: $1.80-$2.50/GPU-hr (rare; SXM on community is uncommon)
• H100 SXM floor: $1.33/GPU-hr interruptible (cheapest published H100)
• H200 verified: $3.50-$5.50/GPU-hr (limited supply)
• B200 verified: $6-$12/GPU-hr (very intermittent availability)
• A100 80GB: $0.79/GPU-hr marketplace floor
• L40S: $0.89/GPU-hr marketplace median
• Reserved: Up to 50% off on-demand with pre-payment

Why Vast.ai delivers the lowest published GPU rates in the category

Vast.ai is a GPU marketplace where individual hosts set their own prices — producing the cheapest published rates at the cost of interruptibility and host variance.

Vast.ai aggregates 40+ data centers worldwide and lets individual hosts rent spare GPU capacity through the marketplace. Prices are not fixed — they move with supply and demand in real-time. On a quiet Tuesday, an H100 PCIe on an unverified host might list at $0.90/hr; on a busy Friday during a training-deadline crunch, the same host might pull the listing entirely. Datacenter-verified hosts offer more stability but push prices to $1.50-$2.27/hr (spheron.network, July 2026).

The platform model is fundamentally different from specialist clouds like Lambda or RunPod Secure. Vast.ai does not own the silicon; it does not guarantee the silicon stays online. For fault-tolerant training with checkpoint discipline and proper host vetting, Vast.ai is the cheapest credible option. For production workloads with strict uptime requirements, hyperscaler on-demand is the safer choice — at 4-8x the price.

H100 pricing on Vast.ai

H100 PCIe is the most common H100 form factor on Vast.ai; H100 SXM listings are rare but command higher rates.

Verified July 2026 (gpucloudcost.com, spheron.network):

• H100 SXM marketplace floor interruptible: $1.33/GPU-hr
• H100 PCIe unverified: $0.90-$1.60/GPU-hr (median ~$1.20)
• H100 PCIe verified datacenter: $1.50-$2.27/GPU-hr (median ~$1.85)
• H100 SXM verified datacenter: $1.80-$2.50/GPU-hr (limited supply)

Comparing to specialist-cloud H100 PCIe: Lambda 1x H100 PCIe at $3.29/GPU-hr; RunPod Community Cloud H100 PCIe at $1.99/GPU-hr; RunPod Secure Cloud H100 PCIe at $2.89/GPU-hr. Vast.ai verified datacenter H100 PCIe at $1.50-$2.27/GPU-hr is 30-50% cheaper than Lambda and competitive with RunPod Community. The tradeoff: Vast.ai reliability is host-level, not platform-level.

For NVLink-dependent workloads (multi-GPU gradient synchronization), H100 SXM is the right chip. Vast.ai SXM verified listings at $1.80-$2.50/GPU-hr are 30-40% cheaper than Lambda 8x SXM at $3.99/GPU-hr, but supply is limited.

H200 pricing on Vast.ai

H200 supply on Vast.ai is thin in 2026 because the bulk of H200 hardware went to hyperscalers and large managed clouds first.

Verified July 2026 (spheron.network):

• H200 verified datacenter: $3.50-$5.50/GPU-hr (median ~$4.50)
• H200 unverified: Occasionally below $3.00/GPU-hr but reliability concerns apply even more acutely because H200 hardware is newer

Compared to RunPod Secure Cloud H200 at $4.59/GPU-hr, Vast.ai verified H200 at $3.50-$5.50/GPU-hr is competitive. Compared to Lambda 1x H200 (not listed on Lambda's pricing card in July 2026), Vast.ai verified is the practical path to single-GPU H200 access.

For H200 workloads, Vast.ai's verified rate is competitive on price but the supply limitations mean capacity planning is harder than with specialist clouds. Reserve capacity before training windows if you commit to Vast.ai for H200.

B200 pricing on Vast.ai

B200 availability on Vast.ai is very limited as of mid-2026 because B200 GPUs require Blackwell-generation HGX hardware.

Verified July 2026 (spheron.network):

• B200 verified datacenter: $6-$12/GPU-hr (median ~$9)
• B200 supply: Intermittent; listings appear and disappear with deployment cycles

Treating Vast.ai as a reliable B200 source would be a planning mistake. The bulk of B200 capacity sits with CoreWeave, Lambda, and RunPod — providers that own or have priority access to Blackwell hardware. For production Blackwell workloads, Vast.ai is at best a supplementary source for opportunistic capacity.

A100, L40S, and older generation pricing

For older-generation GPUs, Vast.ai's marketplace rates are the cheapest published rates in the category.

Verified July 2026 (gpucloudcost.com):

• A100 80GB marketplace floor: $0.79/GPU-hr interruptible
• L40S marketplace median: $0.89/GPU-hr
• RTX 4090 marketplace median: $0.45/GPU-hr (consumer GPU)
• RTX 3090 marketplace floor: $0.21/GPU-hr (consumer GPU)

Compared to RunPod Community Cloud A100 PCIe 40GB at $1.19/GPU-hr and Lambda A100 PCIe 40GB at $1.99/GPU-hr, Vast.ai A100 80GB at $0.79/GPU-hr is 30-50% cheaper. The A100 marketplace on Vast.ai is mature with stable supply — these rates are reliable enough for production training workloads that can tolerate host-level reliability.

Interruptible vs on-demand vs reserved

Vast.ai's three pricing tiers match the major cloud models with marketplace-specific reliability implications.

Verified July 2026 (vast.ai/docs/guides/instances/pricing):

• Interruptible: Lowest cost (50%+ cheaper than on-demand); host can evict at any time; ideal for batch training and fault-tolerant workloads
• On-demand: Fixed pricing; high priority; host cannot actively evict but machine stays online only while host keeps it online; no platform-level uptime guarantee
• Reserved: Discounted rate with pre-payment commitment; up to 50% off on-demand; high priority; preempts interruptible on the same host

The interruptible tier is Vast.ai's signature offering — 50% or more below on-demand pricing. For training workloads with checkpoint discipline (save every 30 minutes or less), interruptible at $1.49/GPU-hr for H100 PCIe is the cheapest credible option in the GPU cloud market.

Side-by-side Vast.ai vs RunPod vs Lambda pricing

How to use Vast.ai safely in 2026

Three rules for using Vast.ai without losing your training job.

1. Use datacenter-verified hosts only. Unverified community hosts can pull hardware offline at any moment. Sort by DLPerf score within the verified host filter for reliability and performance.
2. Checkpoint every 30 minutes or less. Vast.ai interruptible instances can be reclaimed with short notice. Save model state to S3-compatible storage every 30 minutes; checkpointing discipline is the cost of the 50%+ interruptible discount.
3. Verify host reliability before committing compute. Check the host's uptime history, DLPerf score, and prior tenant reviews. A verified host with 95%+ uptime is meaningfully more reliable than one with 70% uptime at the same price.

What enterprise buyers should do next

Three actions for organizations evaluating Vast.ai in 2026.

1. Run a 2-week pilot on interruptible H100 PCIe verified datacenter at $1.50-$2.27/GPU-hr. Same workload as Lambda or RunPod. Compare cost (likely 30-50% lower) against reliability variance (likely higher). Quantify both before committing to a multi-month training workflow.
2. Model storage and egress costs separately. Vast.ai charges for storage per GB-month and host-dependent network transfer. For data-intensive workloads, these line items can add 10-15% to the headline GPU rate.
3. Reserve capacity for time-sensitive workloads. Reserved pricing (up to 50% off on-demand with pre-payment commitment) is the right choice for training jobs with hard deadlines. Interruptible is fine for exploratory work and research; reserved is mandatory for production.

What to watch next

Three near-term datapoints. First, Vast.ai's B200 and Blackwell Ultra listings — supply is currently intermittent; expect more listings as Blackwell hardware proliferates to community data centers by Q1 2027. Second, Vast.ai Serverless pricing — the platform now offers auto-scaling serverless workers with no separate pricing tier, which makes Vast.ai competitive with Modal and Replicate for spiky inference workloads at lower GPU rates. Third, Vast.ai reserved pricing adoption — if reserved capacity scales, Vast.ai's verified datacenter H100 PCIe at $1.50-$2.27/GPU-hr reserved (50% off = $0.75-$1.13/GPU-hr) would undercut every specialist-cloud reserved contract in the market (vast.ai/pricing, July 2026).