t4g instances — General purpose
Burstable Graviton2 general-purpose instances — the cheapest way to run small ARM workloads with burst headroom.
The t4g family sits in AWS's general purpose tier and runs on AWS Graviton (arm64) silicon. It spans7 sizes, from t4g.2xlarge (8 vCPU / 32 GiB) up to t4g.xlarge (4 vCPU /16 GiB). Pricing scales roughly linearly with size, so the per-vCPU cost is similar across the family — pick the smallest size that fits your workload's memory and CPU headroom, then use a Savings Plan or Reserved Instance to cut the effective rate. Graviton families like this one usually deliver the best price/performance in their tier, provided your stack runs on arm64.
All t4g sizes (us-east-1)
| Instance | vCPU | Memory | Network | On-Demand $/hr | Monthly | Max saving |
|---|---|---|---|---|---|---|
| t4g.2xlarge | 8 | 32 GiB | Up to 5 Gigabit | $0.2688 | $196.22 | −62% |
| t4g.large | 2 | 8 GiB | Up to 5 Gigabit | $0.0672 | $49.06 | −62% |
| t4g.medium | 2 | 4 GiB | Up to 5 Gigabit | $0.0336 | $24.53 | −63% |
| t4g.micro | 2 | 1 GiB | Up to 5 Gigabit | $0.0084 | $6.13 | −62% |
| t4g.nano | 2 | 0.5 GiB | Up to 5 Gigabit | $0.0042 | $3.07 | −63% |
| t4g.small | 2 | 2 GiB | Up to 5 Gigabit | $0.0168 | $12.26 | −63% |
| t4g.xlarge | 4 | 16 GiB | Up to 5 Gigabit | $0.1344 | $98.11 | −62% |
List prices, directional. On-Demand, Reserved, and Savings Plan rates are AWS public list prices in USD; effective Savings Plan/RI rates fold in any upfront. Spot is indicative only (derived from the AWS Spot Instance Advisor) — never a firm quote. Always confirm in the AWS console before purchasing. Prices as of 2026-07-01.