Cerebrium

Serverless GPU infrastructure

Founded in 2021, Cerebrium is a serverless infrastructure platform for deploying AI applications. The platform enables developers to deploy large language models, agents, and vision models with automatic scaling from zero to thousands of requests. Cerebrium offers transparent pay-per-second pricing with low cold start times, reducing costs for idle GPU time.

Example customers include Deepgram, Vapi, Tavus, bitHuman.

Cerebrium Homepage

What's good about Cerebrium

Pay-per-second pricing with no cost for idle GPUs
Cold start times as low as a few seconds
Automatic scaling from zero to thousands of requests
Wide range of GPU options from T4 to H200
First 100GB storage free

Cerebrium pricing examples

Cerebrium uses a pay-per-second billing model. You only pay for the compute you use, with pricing based on GPU, CPU, and memory resources.

Below are some example configurations and their estimated costs:

Example configuration	Estimated cost
Block Storage	$5.00 / mo 100 GB beyond free allowance

Cerebrium GPUs

Based on our data, Cerebrium may have GPUs in the following configurations:

Billing

Max price/hr

Min GPUs

Min vCPUs

Min RAM (GB)

of shown

GPU	Total VRAM	vCPUs	RAM	Billing	$/GPU/h	Total/h
1x H100 1x H100 80GB	80GB	16	128GB	On-Demand Billed per second	$2.21	$2.21	View
1x A100 1x A100 40GB	40GB	12	80GB	On-Demand Billed per second	$1.45	$1.45	View
1x H200 1x H200 141GB	141GB	20	180GB	On-Demand Billed per second	$3.30	$3.30	View
1x L40S 1x L40S 48GB	48GB	12	64GB	On-Demand Billed per second	$1.95	$1.95	View
1x L4 1x L4 24GB	24GB	4	24GB	On-Demand Billed per second	$0.80	$0.80	View
1x A10 1x A10 24GB	24GB	8	32GB	On-Demand Billed per second	$1.10	$1.10	View
1x T4 1x T4 16GB	16GB	4	16GB	On-Demand Billed per second	$0.59	$0.59	View
1x A100 1x A100 80GB	80GB	16	120GB	On-Demand Billed per second	$2.06	$2.06	View
No offerings matching your filters.

Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning. You can find Cerebrium's latest pricing here.

Lyceum

Our sponsor

NVIDIA H200 from $2.20/hr, B200 from $3.40/hr, B300 from $3.80/hr

Per-second billing with no minimum commitment or vendor lock-in

Scale to 8,000 GPUs over InfiniBand for distributed AI training

Start now