AWS Instance Comparison

What Does This Tool Compare?

Structural specifications of common EC2 instance families and sizes - vCPU count, memory (GiB), and CPU architecture (x86_64 vs. arm64/Graviton) - side by side, so you can quickly see how instance sizes relate to each other.

Why Use This Tool?

EC2's instance type naming (like t3.medium or m6g.xlarge) encodes family, generation, and size, but doesn't tell you the actual vCPU/memory at a glance. This tool surfaces the real specs for quick comparison.

How to Use It

1. Select two or more instance types from the list.
2. Review their vCPU, memory, and architecture side by side.

Limitations

This reference does not include live, current pricing (which varies by region and purchase option) or every existing instance type/generation - it's a structural specs comparison, not a purchasing tool. Always confirm current details on AWS's own EC2 instance type page.

EC2 Instance Sizes Explained

Within a family, each size step doubles the vCPUs and memory - and roughly the price. For a general purpose M instance (4 GiB of memory per vCPU):

C (compute optimized) instances have half the memory per vCPU (2 GiB) and R (memory optimized) twice as much (8 GiB). Burstable T instances such as t3.micro have 2 vCPUs that run at a baseline level and use CPU credits to burst, which makes them cheap for small websites, test servers and tools that are idle most of the time.

How to Read an Instance Name

In m7g.2xlarge: m is the family (general purpose), 7 the generation, g an extra attribute (Graviton processor) and 2xlarge the size. Common attributes are a (AMD), g (Graviton, Arm), i (Intel), d (local NVMe storage), n (more network bandwidth) and -flex (lower price for workloads that do not need full CPU all the time). Type any name into Explain an instance name above to have it decoded.

Which Instance Should You Choose?

• Small website, blog or test server: a T instance (t3.small, t4g.small) - watch the CPU credits.
• Typical web or API server: M family, starting at large.
• CPU-heavy work such as video encoding, builds or batch jobs: C family.
• Databases and caches that need lots of memory: R family.
• Prefer the newest generation available in your region: it usually gives better performance for the same or lower price.

Graviton, Intel or AMD?

Graviton (Arm) instances, the ones with a g, usually cost less than the matching Intel instance and run most Linux software - including Node.js, Python, PHP, Java and Docker images built for arm64 - without changes. Choose Intel or AMD when you need software that is only available for x86, such as some commercial agents or Windows.