Free cloud & LLM pricing tools — AWS EC2 + LLM API Pricing Explorer

Arm vs x86 savings calculator

Arm chips — AWS Graviton, Azure Cobalt, Google Axion — usually undercut the equal-spec x86 instance on list price. This tool pairs each x86 size with the Arm size that has the exact same vCPU and memory on the same cloud, then computes the saving for a fleet. Only clean equal-spec matches — never an invented equivalence. Free, no signup.

10× t3.2xlarget4g.2xlarge at 730h/mo — save $467.20/mo ($5,606.40/yr)

Option$/hr each10× @ 730h/moAnnual
t3.2xlarge (x86)$0.3328$2,429.44$29,153.28
t4g.2xlarge Arm$0.2688$1,962.24$23,546.88
Difference$0.0640$467.20$5,606.40

List On-Demand prices, 8 vCPU / 32 GiB on both sides. Arm savings assume your workload runs on Arm — verify compatibility first. Prices as of 2026-08-26. Measure your real cloud spend with CloudQuell →

Arm requires workload compatibility

AWS EC2 — x86 vs AWS Graviton

35 equal-spec pairs in us-east-1. Export CSV (free, no account).

Categoryx86AWS Graviton (Arm)vCPU / GiBx86 $/hrArm $/hrSaving
General purposet3.2xlarget4g.2xlarge Arm8 / 32$0.3328$0.2688−19.2%
General purposet3.larget4g.large Arm2 / 8$0.0832$0.0672−19.2%
General purposet3.mediumt4g.medium Arm2 / 4$0.0416$0.0336−19.2%
General purposet3.microt4g.micro Arm2 / 1$0.0104$0.0084−19.2%
General purposet3.nanot4g.nano Arm2 / 0.5$0.0052$0.0042−19.2%
General purposet3.smallt4g.small Arm2 / 2$0.0208$0.0168−19.2%
General purposet3.xlarget4g.xlarge Arm4 / 16$0.1664$0.1344−19.2%
General purposem7i.2xlargem7g.2xlarge Arm8 / 32$0.4032$0.3264−19%
General purposem7i.4xlargem7g.4xlarge Arm16 / 64$0.8064$0.6528−19%
General purposem7i.8xlargem7g.8xlarge Arm32 / 128$1.61$1.31−19%
General purposem7i.12xlargem7g.12xlarge Arm48 / 192$2.42$1.96−19%
General purposem7i.16xlargem7g.16xlarge Arm64 / 256$3.23$2.61−19%
General purposem7i.largem7g.large Arm2 / 8$0.1008$0.0816−19%
General purposem7i.xlargem7g.xlarge Arm4 / 16$0.2016$0.1632−19%
Memory optimizedr7i.2xlarger7g.2xlarge Arm8 / 64$0.5292$0.4284−19%
Memory optimizedr7i.4xlarger7g.4xlarge Arm16 / 128$1.06$0.8568−19%
Memory optimizedr7i.8xlarger7g.8xlarge Arm32 / 256$2.12$1.71−19%
Memory optimizedr7i.12xlarger7g.12xlarge Arm48 / 384$3.18$2.57−19%
Memory optimizedr7i.16xlarger7g.16xlarge Arm64 / 512$4.23$3.43−19%
Memory optimizedr7i.larger7g.large Arm2 / 16$0.1323$0.1071−19%
Memory optimizedr7i.xlarger7g.xlarge Arm4 / 32$0.2646$0.2142−19%
Compute optimizedc7i.2xlargec7g.2xlarge Arm8 / 16$0.3570$0.2900−18.8%
Compute optimizedc7i.4xlargec7g.4xlarge Arm16 / 32$0.7140$0.5800−18.8%
Compute optimizedc7i.8xlargec7g.8xlarge Arm32 / 64$1.43$1.16−18.8%
Compute optimizedc7i.12xlargec7g.12xlarge Arm48 / 96$2.14$1.74−18.8%
Compute optimizedc7i.16xlargec7g.16xlarge Arm64 / 128$2.86$2.32−18.8%
Compute optimizedc7i.largec7g.large Arm2 / 4$0.0893$0.0725−18.8%
Compute optimizedc7i.xlargec7g.xlarge Arm4 / 8$0.1785$0.1450−18.8%
General purposem6i.2xlargem7g.2xlarge Arm8 / 32$0.3840$0.3264−15%
General purposem6i.4xlargem7g.4xlarge Arm16 / 64$0.7680$0.6528−15%
General purposem6i.8xlargem7g.8xlarge Arm32 / 128$1.54$1.31−15%
General purposem6i.12xlargem7g.12xlarge Arm48 / 192$2.30$1.96−15%
General purposem6i.16xlargem7g.16xlarge Arm64 / 256$3.07$2.61−15%
General purposem6i.largem7g.large Arm2 / 8$0.0960$0.0816−15%
General purposem6i.xlargem7g.xlarge Arm4 / 16$0.1920$0.1632−15%

Coverage: 35 x86 sizes mapped to an equal-spec Arm counterpart. Skipped (no exact Arm match): c7i.24xlarge, c7i.48xlarge, m6i.24xlarge, m6i.32xlarge, m7i.24xlarge, m7i.48xlarge, r7i.24xlarge, r7i.48xlarge.

Azure VM — x86 vs Azure Cobalt / Ampere

63 equal-spec pairs in eastus. Export CSV (free, no account).

Categoryx86Azure Cobalt / Ampere (Arm)vCPU / GiBx86 $/hrArm $/hrSaving
General purposeD2as v5D2ps v6 Arm2 / 8$0.1780$0.0702−60.6%
General purposeD48as v5D48ps v6 Arm48 / 192$4.27$1.69−60.6%
General purposeD8as v5D8ps v6 Arm8 / 32$0.7120$0.2810−60.5%
General purposeD16as v5D16ps v6 Arm16 / 64$1.42$0.5620−60.5%
General purposeD4s v6D4ps v6 Arm4 / 16$0.2020$0.1400−30.7%
Memory optimizedE4s v6E4ps v6 Arm4 / 32$0.2650$0.1840−30.6%
General purposeD2s v6D2ps v6 Arm2 / 8$0.1010$0.0702−30.5%
General purposeD32s v6D32ps v6 Arm32 / 128$1.61$1.12−30.4%
General purposeD64s v6D64ps v6 Arm64 / 256$3.23$2.25−30.4%
General purposeD8s v6D8ps v6 Arm8 / 32$0.4030$0.2810−30.3%
General purposeD16s v6D16ps v6 Arm16 / 64$0.8060$0.5620−30.3%
General purposeD48s v6D48ps v6 Arm48 / 192$2.42$1.69−30.3%
General purposeD96s v6D96ps v6 Arm96 / 384$4.84$3.37−30.3%
Memory optimizedE32s v6E32ps v6 Arm32 / 256$2.12$1.48−30.3%
Memory optimizedE48s v6E48ps v6 Arm48 / 384$3.18$2.21−30.3%
Memory optimizedE64s v6E64ps v6 Arm64 / 512$4.23$2.95−30.3%
Memory optimizedE2s v6E2ps v6 Arm2 / 16$0.1320$0.0922−30.2%
Memory optimizedE8s v6E8ps v6 Arm8 / 64$0.5290$0.3690−30.2%
Memory optimizedE16s v6E16ps v6 Arm16 / 128$1.06$0.7380−30.2%
BurstableB2s v2B2ps v2 Arm2 / 8$0.0924$0.0672−27.3%
General purposeD4s v5D4ps v6 Arm4 / 16$0.1920$0.1400−27.1%
Memory optimizedE4s v5E4ps v6 Arm4 / 32$0.2520$0.1840−27%
General purposeD2s v5D2ps v6 Arm2 / 8$0.0960$0.0702−26.9%
General purposeD32s v5D32ps v6 Arm32 / 128$1.54$1.12−26.9%
General purposeD48s v5D48ps v6 Arm48 / 192$2.30$1.69−26.9%
General purposeD64s v5D64ps v6 Arm64 / 256$3.07$2.25−26.9%
General purposeD96s v5D96ps v6 Arm96 / 384$4.61$3.37−26.9%
General purposeD8s v5D8ps v6 Arm8 / 32$0.3840$0.2810−26.8%
General purposeD16s v5D16ps v6 Arm16 / 64$0.7680$0.5620−26.8%
Memory optimizedE2s v5E2ps v6 Arm2 / 16$0.1260$0.0922−26.8%
Memory optimizedE8s v5E8ps v6 Arm8 / 64$0.5040$0.3690−26.8%
Memory optimizedE16s v5E16ps v6 Arm16 / 128$1.01$0.7380−26.8%
Memory optimizedE32s v5E32ps v6 Arm32 / 256$2.02$1.48−26.8%
Memory optimizedE48s v5E48ps v6 Arm48 / 384$3.02$2.21−26.8%
Memory optimizedE64s v5E64ps v6 Arm64 / 512$4.03$2.95−26.8%
Memory optimizedE96s v5E96ps v6 Arm96 / 672$6.05$4.43−26.8%
General purposeD4as v6D4ps v6 Arm4 / 16$0.1820$0.1400−23.1%
General purposeD2as v6D2ps v6 Arm2 / 8$0.0908$0.0702−22.7%
General purposeD32as v6D32ps v6 Arm32 / 128$1.45$1.12−22.7%
General purposeD48as v6D48ps v6 Arm48 / 192$2.18$1.69−22.7%
General purposeD64as v6D64ps v6 Arm64 / 256$2.91$2.25−22.7%
General purposeD96as v6D96ps v6 Arm96 / 384$4.36$3.37−22.7%
Memory optimizedE4as v6E4ps v6 Arm4 / 32$0.2380$0.1840−22.7%
Memory optimizedE32as v6E32ps v6 Arm32 / 256$1.91$1.48−22.7%
Memory optimizedE64as v6E64ps v6 Arm64 / 512$3.81$2.95−22.7%
General purposeD8as v6D8ps v6 Arm8 / 32$0.3630$0.2810−22.6%
General purposeD16as v6D16ps v6 Arm16 / 64$0.7260$0.5620−22.6%
Memory optimizedE8as v6E8ps v6 Arm8 / 64$0.4770$0.3690−22.6%
Memory optimizedE16as v6E16ps v6 Arm16 / 128$0.9540$0.7380−22.6%
Memory optimizedE48as v6E48ps v6 Arm48 / 384$2.86$2.21−22.6%
Memory optimizedE96as v6E96ps v6 Arm96 / 672$5.72$4.43−22.6%
Memory optimizedE2as v6E2ps v6 Arm2 / 16$0.1190$0.0922−22.5%
BurstableB2as v2B2ps v2 Arm2 / 8$0.0844$0.0672−20.4%
BurstableB8as v2B8ps v2 Arm8 / 32$0.3380$0.2690−20.4%
BurstableB8s v2B8ps v2 Arm8 / 32$0.3330$0.2690−19.2%
General purposeD4as v5D4ps v6 Arm4 / 16$0.1720$0.1400−18.6%
General purposeD32as v5D32ps v6 Arm32 / 128$1.38$1.12−18.4%
General purposeD64as v5D64ps v6 Arm64 / 256$2.75$2.25−18.4%
General purposeD96as v5D96ps v6 Arm96 / 384$4.13$3.37−18.4%
BurstableB16s v2B16ps v2 Arm16 / 64$0.7390$0.6110−17.3%
BurstableB4as v2B4ps v2 Arm4 / 16$0.1690$0.1530−9.5%
BurstableB16as v2B16ps v2 Arm16 / 64$0.6750$0.6110−9.5%
BurstableB4s v2B4ps v2 Arm4 / 16$0.1660$0.1530−7.8%

Coverage: 63 x86 sizes mapped to an equal-spec Arm counterpart. Skipped (no exact Arm match): Standard_B32as_v2, Standard_B32s_v2, Standard_D128s_v6, Standard_D192s_v6, Standard_E20as_v6, Standard_E20s_v5, Standard_E20s_v6, Standard_E96s_v6, Standard_E128s_v6. No Arm counterpart series: Falsv6, Fasv6, Fsv2.

GCP Compute — x86 vs Google Axion

84 equal-spec pairs in us-central1. Export CSV (free, no account).

Categoryx86Google Axion (Arm)vCPU / GBx86 $/hrArm $/hrSaving
Compute optimizedc2-standard-4c4a-standard-4 Arm4 / 16$0.2088$0.1796−14%
Compute optimizedc2-standard-8c4a-standard-8 Arm8 / 32$0.4176$0.3592−14%
Compute optimizedc2-standard-16c4a-standard-16 Arm16 / 64$0.8352$0.7184−14%
Compute optimizedc3-highcpu-4c4a-highcpu-4 Arm4 / 8$0.1701$0.1515−10.9%
Compute optimizedc3-highcpu-8c4a-highcpu-8 Arm8 / 16$0.3402$0.3030−10.9%
Compute optimizedc3-highmem-4c4a-highmem-4 Arm4 / 32$0.2646$0.2358−10.9%
Compute optimizedc3-highmem-8c4a-highmem-8 Arm8 / 64$0.5292$0.4715−10.9%
Compute optimizedc3-standard-4c4a-standard-4 Arm4 / 16$0.2016$0.1796−10.9%
Compute optimizedc3-standard-8c4a-standard-8 Arm8 / 32$0.4032$0.3592−10.9%
Compute optimizedc4-highcpu-2c4a-highcpu-2 Arm2 / 4$0.0851$0.0758−10.9%
Compute optimizedc4-highcpu-4c4a-highcpu-4 Arm4 / 8$0.1701$0.1515−10.9%
Compute optimizedc4-highcpu-8c4a-highcpu-8 Arm8 / 16$0.3402$0.3030−10.9%
Compute optimizedc4-highcpu-16c4a-highcpu-16 Arm16 / 32$0.6804$0.6061−10.9%
Compute optimizedc4-highcpu-32c4a-highcpu-32 Arm32 / 64$1.36$1.21−10.9%
Compute optimizedc4-highcpu-48c4a-highcpu-48 Arm48 / 96$2.04$1.82−10.9%
General purposen2-highmem-2c4a-highmem-2 Arm2 / 16$0.1310$0.1179−10%
General purposen2-highmem-4c4a-highmem-4 Arm4 / 32$0.2620$0.2358−10%
General purposen2-highmem-8c4a-highmem-8 Arm8 / 64$0.5241$0.4715−10%
General purposen2-highmem-16c4a-highmem-16 Arm16 / 128$1.05$0.9430−10%
General purposen2-highmem-32c4a-highmem-32 Arm32 / 256$2.10$1.89−10%
General purposen2-standard-2c4a-standard-2 Arm2 / 8$0.0971$0.0898−7.5%
General purposen2-standard-4c4a-standard-4 Arm4 / 16$0.1942$0.1796−7.5%
General purposen2-standard-8c4a-standard-8 Arm8 / 32$0.3885$0.3592−7.5%
General purposen2-standard-16c4a-standard-16 Arm16 / 64$0.7769$0.7184−7.5%
General purposen2-standard-32c4a-standard-32 Arm32 / 128$1.55$1.44−7.5%
Compute optimizedc2d-highmem-2c4a-highmem-2 Arm2 / 16$0.1225$0.1179−3.7%
Compute optimizedc2d-highmem-4c4a-highmem-4 Arm4 / 32$0.2449$0.2358−3.7%
Compute optimizedc2d-highmem-8c4a-highmem-8 Arm8 / 64$0.4899$0.4715−3.7%
Compute optimizedc2d-highmem-16c4a-highmem-16 Arm16 / 128$0.9798$0.9430−3.7%
Compute optimizedc2d-highmem-32c4a-highmem-32 Arm32 / 256$1.96$1.89−3.7%
Compute optimizedc3d-highmem-4c4a-highmem-4 Arm4 / 32$0.2449$0.2358−3.7%
Compute optimizedc3d-highmem-8c4a-highmem-8 Arm8 / 64$0.4899$0.4715−3.7%
Compute optimizedc3d-highmem-16c4a-highmem-16 Arm16 / 128$0.9798$0.9430−3.7%
Compute optimizedc2d-standard-2c4a-standard-2 Arm2 / 8$0.0908$0.0898−1.1%
Compute optimizedc2d-standard-4c4a-standard-4 Arm4 / 16$0.1816$0.1796−1.1%
Compute optimizedc2d-standard-8c4a-standard-8 Arm8 / 32$0.3632$0.3592−1.1%
Compute optimizedc2d-standard-16c4a-standard-16 Arm16 / 64$0.7264$0.7184−1.1%
Compute optimizedc2d-standard-32c4a-standard-32 Arm32 / 128$1.45$1.44−1.1%
Compute optimizedc3d-standard-4c4a-standard-4 Arm4 / 16$0.1816$0.1796−1.1%
Compute optimizedc3d-standard-8c4a-standard-8 Arm8 / 32$0.3632$0.3592−1.1%
Compute optimizedc3d-standard-16c4a-standard-16 Arm16 / 64$0.7264$0.7184−1.1%
General purposen4-highcpu-2c4a-highcpu-2 Arm2 / 4$0.0765$0.0758−1%
General purposen4-highcpu-4c4a-highcpu-4 Arm4 / 8$0.1531$0.1515−1%
General purposen4-highcpu-8c4a-highcpu-8 Arm8 / 16$0.3062$0.3030−1%
General purposen4-highcpu-16c4a-highcpu-16 Arm16 / 32$0.6123$0.6061−1%
General purposen4-highcpu-32c4a-highcpu-32 Arm32 / 64$1.22$1.21−1%
General purposen4-highcpu-48c4a-highcpu-48 Arm48 / 96$1.84$1.82−1%
General purposen4-highcpu-64c4a-highcpu-64 Arm64 / 128$2.45$2.42−1%
General purposen4-highmem-2c4a-highmem-2 Arm2 / 16$0.1190$0.1179−1%
General purposen4-highmem-4c4a-highmem-4 Arm4 / 32$0.2380$0.2358−1%
General purposen4-highmem-8c4a-highmem-8 Arm8 / 64$0.4761$0.4715−1%
General purposen4-highmem-16c4a-highmem-16 Arm16 / 128$0.9522$0.9430−1%
General purposen4-highmem-32c4a-highmem-32 Arm32 / 256$1.90$1.89−1%
General purposen4-highmem-48c4a-highmem-48 Arm48 / 384$2.86$2.83−1%
General purposen4-highmem-64c4a-highmem-64 Arm64 / 512$3.81$3.77−1%
General purposen4-standard-2c4a-standard-2 Arm2 / 8$0.0907$0.0898−1%
General purposen4-standard-4c4a-standard-4 Arm4 / 16$0.1814$0.1796−1%
General purposen4-standard-8c4a-standard-8 Arm8 / 32$0.3628$0.3592−1%
General purposen4-standard-16c4a-standard-16 Arm16 / 64$0.7256$0.7184−1%
General purposen4-standard-32c4a-standard-32 Arm32 / 128$1.45$1.44−1%
General purposen4-standard-48c4a-standard-48 Arm48 / 192$2.18$2.16−1%
General purposen4-standard-64c4a-standard-64 Arm64 / 256$2.90$2.87−1%
Compute optimizedc2d-highcpu-2c4a-highcpu-2 Arm2 / 4$0.0750$0.0758−-1.1%
Compute optimizedc2d-highcpu-4c4a-highcpu-4 Arm4 / 8$0.1499$0.1515−-1.1%
Compute optimizedc2d-highcpu-8c4a-highcpu-8 Arm8 / 16$0.2998$0.3030−-1.1%
Compute optimizedc2d-highcpu-16c4a-highcpu-16 Arm16 / 32$0.5997$0.6061−-1.1%
Compute optimizedc2d-highcpu-32c4a-highcpu-32 Arm32 / 64$1.20$1.21−-1.1%
Compute optimizedc3d-highcpu-4c4a-highcpu-4 Arm4 / 8$0.1499$0.1515−-1.1%
Compute optimizedc3d-highcpu-8c4a-highcpu-8 Arm8 / 16$0.2998$0.3030−-1.1%
Compute optimizedc3d-highcpu-16c4a-highcpu-16 Arm16 / 32$0.5997$0.6061−-1.1%
General purposen2d-highmem-2c4a-highmem-2 Arm2 / 16$0.1140$0.1179−-3.4%
General purposen2d-highmem-4c4a-highmem-4 Arm4 / 32$0.2280$0.2358−-3.4%
General purposen2d-highmem-8c4a-highmem-8 Arm8 / 64$0.4559$0.4715−-3.4%
General purposen2d-highmem-16c4a-highmem-16 Arm16 / 128$0.9118$0.9430−-3.4%
General purposen2d-highmem-32c4a-highmem-32 Arm32 / 256$1.82$1.89−-3.4%
General purposen2d-highmem-48c4a-highmem-48 Arm48 / 384$2.74$2.83−-3.4%
General purposen2d-highmem-64c4a-highmem-64 Arm64 / 512$3.65$3.77−-3.4%
General purposen2d-standard-2c4a-standard-2 Arm2 / 8$0.0845$0.0898−-6.3%
General purposen2d-standard-4c4a-standard-4 Arm4 / 16$0.1690$0.1796−-6.3%
General purposen2d-standard-8c4a-standard-8 Arm8 / 32$0.3380$0.3592−-6.3%
General purposen2d-standard-16c4a-standard-16 Arm16 / 64$0.6759$0.7184−-6.3%
General purposen2d-standard-32c4a-standard-32 Arm32 / 128$1.35$1.44−-6.3%
General purposen2d-standard-48c4a-standard-48 Arm48 / 192$2.03$2.16−-6.3%
General purposen2d-standard-64c4a-standard-64 Arm64 / 256$2.70$2.87−-6.3%

Coverage: 84 x86 sizes mapped to an equal-spec Arm counterpart. Skipped (no exact Arm match): c2-standard-30, c2-standard-60, c2d-highcpu-56, c2d-highcpu-112, c2d-highmem-56, c2d-highmem-112, c2d-standard-56, c2d-standard-112, c3-highcpu-22, c3-highcpu-44, c3-highcpu-88, c3-highcpu-176, c3-highmem-22, c3-highmem-44, c3-highmem-88, c3-highmem-176, c3-standard-22, c3-standard-44, c3-standard-88, c3-standard-176, c3d-highcpu-30, c3d-highcpu-60, c3d-highcpu-90, c3d-highcpu-180, c3d-highcpu-360, c3d-highmem-30, c3d-highmem-60, c3d-highmem-90, c3d-highmem-180, c3d-highmem-360, c3d-standard-30, c3d-standard-60, c3d-standard-90, c3d-standard-180, c3d-standard-360, c4-highcpu-96, c4-highcpu-192, c4-highmem-2, c4-highmem-4, c4-highmem-8, c4-highmem-16, c4-highmem-32, c4-highmem-48, c4-highmem-96, c4-highmem-192, c4-standard-2, c4-standard-4, c4-standard-8, c4-standard-16, c4-standard-32, c4-standard-48, c4-standard-96, n2-highcpu-2, n2-highcpu-4, n2-highcpu-8, n2-highcpu-16, n2-highcpu-32, n2d-highcpu-2, n2d-highcpu-4, n2d-highcpu-8, n2d-highcpu-16, n2d-highcpu-32, n2d-highcpu-48, n2d-highcpu-64, n2d-highcpu-80, n2d-highcpu-96, n2d-highcpu-128, n2d-highcpu-224, n2d-highmem-80, n2d-highmem-96, n2d-standard-80, n2d-standard-96, n2d-standard-128, n2d-standard-224, n4-highcpu-80, n4-highmem-80, n4-standard-80. No Arm counterpart series: e2, m3.

From list prices to your real spend

This calculator uses list prices and assumes a like-for-like move. CloudQuell measures your real cloud spend and commitment coverage across AWS, Azure and GCP, so an x86→Arm move is weighed against your actual usage — not a slider.

Explore the CloudQuell platform →

Related: Savings Plan break-even →EC2 Explorer →What is Axion? →

List prices, directional. Arm savings are computed from public On-Demand list prices in USD for equal-spec (identical vCPU + memory) instance pairs. Realising them requires ARM64 workload compatibility — treat this as planning, not a quote. Prices as of 2026-08-26.