GPU cloud · head to head

Modal vs TensorDock

What each costs, whether there's a free plan, and what each is best for, from prices checked on their own sites.

Modal and TensorDock at a glance
Fact	Modal	TensorDock
From	$1.10/hr	$0.35/hr
Free plan	Yes · $30/month of free credits on the Starter plan	No · No ($5 minimum deposit)
Best for	Serverless GPUs for Python: pay per second while your code runs.	A GPU cloud marketplace with low hourly prices on consumer and data-centre GPUs.
Checked	Oct 17, 2026	Oct 17, 2026
Website	Visit	Visit

The difference in short

Modal plans and prices
Plan	Price	Specs / includes	Notes
B200	$6.25/hr	—	Published only per second: $0.001736/sec. Hourly = x3600, exact arithmetic. GPU price only; CPU and memory billed separately. VRAM not stated on the pricing page.
H200 SXM	$4.54/hr	—	Published $0.001261/sec; hourly = x3600 exactly. GPU only. VRAM not stated on the pricing page.
H100 SXM5	$3.95/hr	—	Published $0.001097/sec; hourly = x3600 exactly. GPU only. H100 is 80GB, but the pricing page does not state VRAM.
A100 80GB	$2.50/hr	—	Published $0.000694/sec; hourly = x3600 exactly. A100 40GB is $0.000583/sec = $2.0988/hr.
L40S	$1.95/hr	—	Published $0.000542/sec; hourly = x3600 exactly. L40S is 48GB, but the pricing page does not state VRAM.
A10	$1.10/hr	—	Published $0.000306/sec; hourly = x3600 exactly. Also listed: L4 $0.000222/sec ($0.7992/hr), T4 $0.000164/sec ($0.5904/hr), RTX PRO 6000 $0.000842/sec ($3.0312/hr), B300 $0.001972/sec ($7.0992/hr).

TensorDock plans and prices
Plan	Price	Specs / includes	Notes
H100 80GB SXM5	$2.25/hr	H100 SXM5	Marketplace, varies. Homepage card 'H100 SXM5 FROM $2.25/HR'; homepage banner (undated, 'March News') also says on-demand H100 SXM at $2.25/hr. vCPU/RAM are chosen by the customer.
A100 80GB SXM4	$1.80/hr	A100 SXM4	Marketplace, varies. Homepage card 'A100 SXM4 FROM $1.80/HR'; the 80GB size comes from tensordock.com/cloud-gpus.html (table marked 'Last Updated: July 24, 2024'), which also lists A100 PCIe 80GB at $1.50/hr.
RTX 4090 24GB	$0.35/hr	—	Marketplace, varies. Homepage card 'RTX 4090 FROM $0.35/HR'. Consumer GPUs advertised from $0.12/hr (model not named).

Cheapest first where the prices compare.

Which is cheaper, Modal or TensorDock?

TensorDock costs less to start: $0.35/hr against $1.10/hr for Modal. Prices checked Oct 17, 2026 on each company's own site.

Do Modal and TensorDock have a free plan?

Modal: Yes. $30/month of free credits on the Starter plan. TensorDock: No. $5 minimum deposit.

What else should I compare with Modal and TensorDock?

Other GPU cloud options we've checked: RunPod ($0.34/hr), Vast.ai ($0.49/hr), DigitalOcean GPU Droplets ($0.76/hr), Hyperstack ($1.35/hr).

Built from the prices and facts we checked on each company's own site. We haven't tested either hands-on.How we work