Vast.ai vs Modal

Round by round

1. Round 1The cheapest H100Vast.ai wins
   Vast.ai8/10

   $1.48

   H100 SXM

   • the lowest H100 offer on the marketplace (offers seen ran from about $2 to $19)

   Source: vast.ai ↗

   Modal5/10

   $3.95

   H100 at $0.001097 a second, about $3.95 an hour

   • $3.95, against $0.90

   Source: modal.com ↗

   Vast.ai wins this round on what the company itself publishes.

2. Round 2The cheapest A100Vast.ai wins
   Vast.ai8/10

   $0.37/hr

   A100 PCIE GPU price

   • the lowest A100 PCIe offer on the marketplace

   Source: vast.ai ↗

   Modal5/10

   $2.10

   A100 40GB at $0.000583 a second, about $2.10 an hour

   • $2.10, against $0.35

   Source: modal.com ↗

   Vast.ai wins this round on what the company itself publishes.

3. Round 3How cheap you can startVast.ai wins
   Vast.ai8/10

   $0.11

   RTX 3090

   • the lowest rate on the marketplace, for consumer GPUs

   Source: vast.ai ↗

   Modal5/10

   $0.59

   a T4 at $0.000164 a second

   • $0.59, against $0.13

   Source: modal.com ↗

   Vast.ai wins this round on what the company itself publishes.

4. Round 4What it costs to get your data outModal wins
   Vast.ai5/10

   Bandwidth charged both ways

   billed per byte sent or received, at rates that vary by host

   • Bandwidth charged both ways, against 1 TiB free, then $0.04 per GiB

   Source: docs.vast.ai ↗

   Modal8/10

   1 TiB free, then $0.04 per GiB

   network egress is billed from 1 October 2026; the Starter plan includes 1 TiB a cycle

   Source: modal.com ↗

   Modal wins this round on what the company itself publishes.

5. Round 5How precisely you are billedTie
   Vast.ai7/10

   Per-second billing

   at raw marketplace rates

   Source: vast.ai ↗

   Modal7/10

   Billed by the second

   with nothing charged for idle time

   Source: modal.com ↗

   Nothing to separate them here.

6. Round 6Whether GPUs can run serverlessTie
   Vast.ai7/10

   Yes

   marketplace GPU cloud (on-demand / interruptible / reserved)

   Source: vast.ai ↗

   Modal7/10

   Yes

   serverless pay-per-second GPU pricing

   Source: modal.com ↗

   Nothing to separate them here.

Side by side

Best for the cheapest h100Vast.ai

Check price

Modal

Check price

Overall score

7.2

Best

6.4

Price

See price

See price

The cheapest H100

8.0

6.6

The cheapest A100

8.0

6.2

What it costs to get your data out

5.3

6.6

How many places it runs

8.0

—

How cheap you can start

7.4

6.0

Head-to-head findings

The cheapest H100

$1.73

$3.95

The cheapest A100

$0.37/hr

$2.10

What it costs to get your data out

Bandwidth charged both ways

1 TiB free, then $0.04 per GiB

How many places it runs

40+

—

How cheap you can start

$0.11/hr

$0.59

What customers say on Trustpilot

4.1 out of 5

—

How precisely you are billed

Per-second billing

Billed by the second

What storage costs per month

—

$0.09/GiB-mo (Volumes)

Record

Head-to-head record

8–0

3–2

Head-to-heads won

100 %

60 %

Rounds won

65 %

35 %

Pros

• the lowest H100 offer on the marketplace (offers seen ran from about $2 to $19)
• the lowest A100 PCIe offer on the marketplace
• the lowest rate on the marketplace, for consumer GPUs
• host locations on a marketplace, not data-center regions

Cons

• Bandwidth charged both ways, against No egress fees
• Bandwidth charged both ways, against Free
• Up to 50% off, against Up to 60% off
• $0.13, against Free GPU notebooks

• $3.95, against $1.99
• $2.10, against $1.19
• $4.54, against $3.59
• $6.25, against $5.98

More cloud GPU providers match-ups