Replicate vs TensorWave

Replicate's Pricing

Replicate uses a metered billing model. You only pay as long as your code is running, billed by the second based on the hardware selected.

Replicate GPUs

Name	GPUs	VRAM	vCPUs	RAM	Price/h
1x Nvidia T4	1x T4	16GB	4	16GB	$0.81	Source
1x Nvidia L40S	1x L40S	48GB	10	65GB	$3.51	Source
1x Nvidia A100 (80GB)	1x A100	80GB	10	144GB	$5.04	Source
1x Nvidia H100	1x H100	80GB	13	72GB	$5.49	Source
2x Nvidia L40S	2x L40S	96GB	20	144GB	$7.02	Source
2x Nvidia A100 (80GB)	2x A100	160GB	20	288GB	$10.08	Source
2x Nvidia H100	2x H100	160GB	26	144GB	$10.98	Source
4x Nvidia A100 (80GB)	4x A100	320GB	40	576GB	$20.16	Source
4x Nvidia H100	4x H100	320GB	52	288GB	$21.96	Source
8x Nvidia A100 (80GB)	8x A100	640GB	80	960GB	$40.32	Source
8x Nvidia H100	8x H100	640GB	104	576GB	$43.92	Source
Show 4 more...

TensorWave's Pricing

TensorWave offers bare metal GPU nodes and fully managed Kubernetes clusters. You can reserve nodes for periods ranging from 6 months to 3 years.

While their pricing is not yet publicly available, you can request a quote on their website.

TensorWave GPUs

Name	GPUs	VRAM	vCPUs	RAM	Price/h
AMD MI300X	8x MI300X	1536GB	--	--	On Request	Source

Here are some managed services that Replicate and TensorWave offer:

Service	Replicate	TensorWave
GPU-powered Servers

Want to see how Replicate and TensorWave compare against other providers? Check out these other comparisons: