Instance grid · atlas

Each row is an instance family, each column a size. One step right doubles the vCPUs, the memory and the price, so the cheapest family for your workload is the lightest row.

Cheapest for its sizePriciest

General purpose 4 GiB memory per vCPU
	medium 1 vCPU	large 2 vCPU	xlarge 4 vCPU	2xlarge 8 vCPU	4xlarge 16 vCPU	8xlarge 32 vCPU	12xlarge 48 vCPU	16xlarge 64 vCPU	24xlarge 96 vCPU	32xlarge 128 vCPU	48xlarge 192 vCPU	metal whole host
m7g Graviton3	$23	$46	$92	$184	$369	$738	$1.1k	$1.5k				$1.5k
m7i Xeon Sapphire Rapids		$57	$114	$228	$456	$911	$1.4k	$1.8k	$2.7k		$5.5k
m6g Graviton2	$21	$41	$82	$164	$328	$657	$985	$1.3k				$1.3k
m6gd Graviton2	$24	$49	$97	$195	$389	$779	$1.2k	$1.6k				$1.6k
m6i Xeon Ice Lake		$54	$109	$217	$434	$868	$1.3k	$1.7k	$2.6k	$3.5k		$3.5k
m6id Xeon Ice Lake		$64	$128	$256	$511	$1.0k	$1.5k	$2.0k	$3.1k	$4.1k		$4.1k
m5 Xeon Skylake Newest		$52	$104	$207	$413	$826	$1.2k	$1.7k	$2.5k			$2.5k
m5d Xeon Skylake		$61	$122	$243	$487	$974	$1.5k	$1.9k	$2.9k			$2.9k
Compute optimized 2 GiB memory per vCPU
	medium 1 vCPU	large 2 vCPU	xlarge 4 vCPU	2xlarge 8 vCPU	4xlarge 16 vCPU	8xlarge 32 vCPU	12xlarge 48 vCPU	16xlarge 64 vCPU	24xlarge 96 vCPU	32xlarge 128 vCPU	48xlarge 192 vCPU	metal whole host
c8g Graviton4	$22	$43	$87	$173	$346	$692	$1.0k	$1.4k	$2.1k		$4.2k
c7g Graviton3	$20	$39	$79	$157	$315	$629	$944	$1.3k				$1.3k
c7i Xeon Sapphire Rapids		$49	$97	$194	$389	$777	$1.2k	$1.6k	$2.3k		$4.7k
c6g Graviton2	$19	$37	$74	$148	$296	$592	$888	$1.2k				$1.2k
c6gn Graviton2	$24	$47	$94	$188	$376	$753	$1.1k	$1.5k
c6i Xeon Ice Lake		$46	$93	$185	$370	$740	$1.1k	$1.5k	$2.2k	$3.0k		$3.0k
c6id Xeon Ice Lake		$55	$111	$221	$442	$884	$1.3k	$1.8k	$2.7k	$3.5k		$3.5k
c6in Xeon Ice Lake		$62	$124	$247	$495	$990	$1.5k	$2.0k	$3.0k	$4.0k		$4.0k
c5 Xeon Skylake Newest		$47	$93	$185	$371		$1.1k		$2.2k			$2.2k
c5d Xeon Skylake		$53	$105	$210	$421		$1.3k		$2.5k			$2.5k
Memory optimized 8 GiB or more memory per vCPU
	medium 1 vCPU	large 2 vCPU	xlarge 4 vCPU	2xlarge 8 vCPU	4xlarge 16 vCPU	8xlarge 32 vCPU	12xlarge 48 vCPU	16xlarge 64 vCPU	24xlarge 96 vCPU	32xlarge 128 vCPU	48xlarge 192 vCPU	metal whole host
r7g Graviton3	$30	$61	$122	$243	$486	$972	$1.5k	$1.9k				$1.9k
r7gd Graviton3	$36	$73	$146	$292	$584	$1.2k	$1.8k	$2.3k				$2.3k
r6g Graviton2	$27	$54	$109	$218	$436	$872	$1.3k	$1.7k				$1.7k
r6i Xeon Ice Lake		$71	$143	$286	$572	$1.1k	$1.7k	$2.3k	$3.4k	$4.6k		$4.6k
r6id Xeon Ice Lake		$81	$162	$325	$649	$1.3k	$1.9k	$2.6k	$3.9k	$5.2k		$5.2k
r5 Xeon Skylake Newest		$68	$137	$272	$545	$1.1k	$1.6k	$2.2k	$3.3k			$3.3k
r5d Xeon Skylake		$77	$155	$309	$618	$1.2k	$1.9k	$2.5k	$3.7k			$3.7k
x2idn Xeon Ice Lake								$3.8k	$5.7k	$7.5k		$7.5k
Storage optimized Fast local disks attached to the host
	medium 1 vCPU	large 2 vCPU	xlarge 4 vCPU	2xlarge 8 vCPU	4xlarge 16 vCPU	8xlarge 32 vCPU	12xlarge 48 vCPU	16xlarge 64 vCPU	24xlarge 96 vCPU	32xlarge 128 vCPU	48xlarge 192 vCPU	metal whole host
i4i Xeon Ice Lake		$95	$189	$378	$756	$1.5k		$3.0k		$6.1k		$6.1k
d3 Xeon Cascade Lake Newest			$294	$588	$1.2k	$2.4k
i3 Xeon Broadwell Newest		$90	$180	$361	$721	$1.4k		$2.9k
i3en Xeon Skylake		$131	$262	$523			$3.1k		$6.3k			$6.3k
Accelerated computing GPUs and AWS ML chips
	medium 1 vCPU	large 2 vCPU	xlarge 4 vCPU	2xlarge 8 vCPU	4xlarge 16 vCPU	8xlarge 32 vCPU	12xlarge 48 vCPU	16xlarge 64 vCPU	24xlarge 96 vCPU	32xlarge 128 vCPU	48xlarge 192 vCPU	metal whole host
g5 EPYC Rome · 24 GB per GPU			$542	$653	$875	$1.3k	$3.1k ×4	$2.2k	$4.4k ×4		$8.8k ×8
p4de Xeon Cascade Lake · 80 GB per GPU									$16k ×8
Burstable Cheap baseline CPU that bursts on credits
	nano	micro	small	medium	large	xlarge	2xlarge
t4g Graviton2	$2.19	$4.45	$8.91	$18	$36	$71	$142
t3 Xeon Skylake Newest	$2.77	$5.47	$11	$22	$44	$88	$176
t3a EPYC Naples	$2.48	$4.96	$9.86	$20	$39	$79	$158