Compare instances

4 instances in Tokyo, on-demand prices.

Saves $1,553 a year over the most expensive, at 1 instance.

Graviton4, +10% vs the cheapest.

	m7g.2xlarge × Graviton3 · 8 vCPU · 32 GiB $308/mo $3,693 a year for 1	m8g.2xlarge × Graviton4 · 8 vCPU · 32 GiB $339/mo $4,063 a year for 1	m7i.2xlarge × Xeon Sapphire Rapids · 8 vCPU · 32 GiB $380/mo $4,562 a year for 1	m7a.2xlarge × EPYC Genoa · 8 vCPU · 32 GiB $437/mo $5,247 a year for 1
Price
On-demand	$308 BEST	$339 +10%	$380 +24%	$437 +42%
Reserved 1 yr	$204 BEST	$224 +10%	$252 +24%	$290 +42%
Reserved 3 yr	$139 BEST	$153 +10%	$172 +24%	$198 +42%
Per vCPU	$38.47 BEST	$42.32 +10%	$47.52 +24%	$54.65 +42%
Per GiB of memory	$9.618 BEST	$10.58 +10%	$11.88 +24%	$13.66 +42%
Compute
Processor	Graviton3	Graviton4	Xeon Sapphire Rapids	EPYC Genoa
Clock speed	2.5 GHz −22%	2.7 GHz −16%	3.2 GHz BEST	—
Architecture	arm64	arm64	x86_64	x86_64
Generation	Gen 7	Gen 8	Gen 7	Gen 7
vCPUs	8	8	8	8
Memory & storage
Memory	32 GiB	32 GiB	32 GiB	32 GiB
Memory per vCPU	4 GiB	4 GiB	4 GiB	4 GiB
Local disk	EBS only	EBS only	EBS only	EBS only
Network
Network	Up to 15 Gbps BEST	Up to 15 Gbps BEST	Up to 12.5 Gbps −17%	Up to 12.5 Gbps −17%
EBS bandwidth	Up to 10 Gbps	Up to 10 Gbps	Up to 10 Gbps	Up to 10 Gbps