AWS Inf1 instance types

Inf1 is a gpu & accelerated Amazon EC2 family with 4 instance types, from 4 – 96 vCPUs and 8 GiB – 192 GiB of memory with AWS Inferentia accelerators, running on Intel Xeon Platinum 8275CL (Cascade Lake). Linux on demand starts at $0.228 per hour.

gpu & acceleratedx86-64
sizes
4
vCPUs
4 – 96
memory
8 GiB – 192 GiB
accelerator
Inferentia
on-demand from
$0.228

per hour · 22 regions

Inf1 sizes and prices

Linux, hourly, US East (N. Virginia) (other regions on each size's page)

SizevCPUMemoryAcceleratorLocal storageOn-demandMonthlySpot1-yr
inf1.xlarge48 GiB1 × AWS Inferentia–$0.228$166$0.0908$0.144
inf1.2xlarge816 GiB1 × AWS Inferentia–$0.362$264$0.1695$0.228
inf1.6xlarge2448 GiB4 × AWS Inferentia–$1.18$861$0.5169$0.744
inf1.24xlarge96192 GiB16 × AWS Inferentia–$4.721$3,446$1.968$2.974

Other gpu & accelerated families on AWS

Inf1 FAQ

What sizes does the AWS Inf1 family have?

Inf1 has 4 instance types, from inf1.xlarge (4 vCPU, 8 GiB) to inf1.24xlarge (96 vCPU, 192 GiB). They run on Intel Xeon Platinum 8275CL (Cascade Lake).

What is the cheapest Inf1 instance type?

inf1.xlarge at $0.228 per hour (Linux, on demand) in US East (N. Virginia), about $166 per month.

What is Inf1 best for?

Inf1 is a gpu & accelerated family. GPUs and AI accelerators for training, inference, rendering and video.