Quick Answer
Last verified:
Medium confidence

Vast.ai costs $0.02 to $3.85 per GPU/hour as of July 2026, with 3 plans available. Plans: Interruptible at $0.1/GPU/hour, On-Demand at $0.3/GPU/hour, and Reserved at $0.2/GPU/hour. Pricing depends on your chosen tier, contract length, and negotiated discounts.

Use the interactive pricing calculator to estimate your exact cost based on team size and requirements.

  • Free tier: No free tier available

Vast.ai offers 3 pricing tiers: Interruptible, On-Demand, Reserved. Paid plans include Interruptible at $0.1/user/month, On-Demand at $0.3/user/month, Reserved at $0.2/user/month. The On-Demand plan is production workloads requiring guaranteed uptime.

Last verified: | Sources: 2 pricing pages and reviews
Know Before You Buy — See Vast.ai Plans

Vast.ai lists $0.02-$3.85/GPU/hour, but hidden costs like implementation and support add to the total as of July 2026. Key hidden costs: serverless inference pricing premium vs. api alternatives, interruptible instance interruption loss, bandwidth / egress charges. Verified from 2 sources by CostBench.

Hidden Costs Breakdown

1

Serverless Inference Pricing Premium vs. API Alternatives

high addon

Vast.ai's serverless per-minute AI model inference is significantly more expensive than dedicated API pricing from providers like OpenRouter. Community users report paying $0.30/minute ($18/hour) for serverless model access, which can far exceed the cost of equivalent API-based inference for common models.

hn

vast offers serverless per minute AI models so checking it for something like https://vast.ai/model/deepseek-v3.2-exp or even glm 3.6 basically every of these turns out to be $.30 cents/minute or 18$ per hour

2

Interruptible Instance Interruption Loss

medium overage

The lowest-cost Interruptible tier can be preempted at any time by higher-priority bids. Work that has not been checkpointed is lost, and users effectively pay for compute time that does not complete. Workloads without robust checkpointing can see significant wasted spend.

hn

Your pricing is around 50-100% higher than vast.ai, a relatively mature competitor in this space who offers the same GPUs and more with beefier rigs.

3

Bandwidth / Egress Charges

high overage

Vast.ai charges separately for internet bandwidth usage on top of the hourly GPU rate. Users report being billed significant amounts for data transfer even during short sessions.

trustpilot

On top of the price/h of your specific server you will be also charged ungodly amounts for ***internet bandwidth*** in my case it was $2.5/100gb. keep in mind that i only used the instance for 20 minutes.

4

Billing Continues After Instance Becomes Inaccessible

critical overage

Multiple users report being charged for hours or days when their instance went offline or became unreachable, with no automatic billing pause and support unwilling to issue refunds for downtime.

trustpilot

I rented a high-end GPU server (RTX 3090 Ti) on Vast.ai with strong specs and a price around $0.143/hour (~0.54 SAR/hour). However, the service is practically unusable.

trustpilot

Vast.ai continued to charge my account, even though I had zero access to the server. This is unacceptable for a service that claims to be reliable and user-friendly.

5

Charges for Failed or Non-Functional Instances

high overage

Users report being billed for instances that fail to start, never load, or experience degraded performance — with support often refusing refunds.

trustpilot

I was charged $1.70 for ~20 minutes of completely non-functional service. Support claimed the charge was only $1.20 for a model download, which is an unacceptable excuse for downtime and unfair billing.

trustpilot

Many of the lowest-priced instances are full of bugs and issues. In the end, those supposedly cheap options don’t function properly, and the ones that do work are the more expensive ones.

6

GPU Performance Degradation Without Price Adjustment

high addon

Verified hosts have been reported to silently reduce GPU performance after rental begins, with no corresponding price reduction or notification to the renter.

trustpilot

I discovered the host had silently reduced the GPU performance by 22% with no notice and no price adjustment — I was paying full price for a degraded machine.

trustpilot

Rented instances removed GPUs while still charging the full price. They restarted my instances without warning on multiple occasions.

7

Trial-and-Error Cost from Unreliable Instances

medium overage

Users commonly need to rent and discard multiple instances before finding one that works reliably, incurring small but real costs each attempt.

trustpilot

I had to rent five different instances before finding one that actually worked, which was a waste of both time and money (not a huge amount). It’s honestly hair-pulling frustrating.

8

Data Loss Risk from Sudden Instance Termination

critical migration

Instances are terminated without warning, and data stored on them may become inaccessible or lost. Recovering data to another instance incurs additional bandwidth charges.

trustpilot

your instance gpus can get scheduled from another guy and you have to copy data which costs bandwidth to another instance just so you can open another even if its targeted as "on demand"

trustpilot

My data was stuck on a machine I couldn’t access, my time was wasted for more than 24 hours, and the overall stress caused by their unresponsive support was simply terrible.

Frequently Asked Questions

01 What hidden costs should I budget for with Vast.ai?

Beyond the license fee, budget for: Serverless Inference Pricing Premium vs. API Alternatives ($18/hr); Interruptible Instance Interruption Loss (10-30% of license costs); Bandwidth / Egress Charges ($2.50/100GB); Billing Continues After Instance Becomes Inaccessible (5-25% of total spend); Charges for Failed or Non-Functional Instances (5-15% of license costs); GPU Performance Degradation Without Price Adjustment (10-25% effective cost increase); Trial-and-Error Cost from Unreliable Instances (5-20% of total spend); Data Loss Risk from Sudden Instance Termination (5-15% of license costs). Exact totals depend on your deployment size and negotiated terms.

02 Does Vast.ai charge for implementation?

Implementation costs for Vast.ai vary by deployment size and customization. Contact the vendor or check our sourced hidden-cost breakdown above for verified figures.

03 How much does Vast.ai support cost?

Premium support pricing for Vast.ai depends on your tier and contract terms. See the sourced cost breakdown above for any verified figures we have.

04 Are there overage or storage costs with Vast.ai?

The lowest-cost Interruptible tier can be preempted at any time by higher-priority bids. Work that has not been checkpointed is lost, and users effectively pay for compute time that does not complete. Estimated impact: 10-30% of license costs.

05 What add-ons cost extra with Vast.ai?

Add-on pricing for Vast.ai varies by feature. The sourced cost breakdown above lists any verified add-on costs we have.

Now You Know the Real Cost — See Vast.ai Plans

Check current Vast.ai pricing

Prices and terms change; verify against the live pricing page.

See Vast.ai Pricing