AWS T3 Vs. T3a Vs. T4g Instances: A Side-By-Side Comparison

A unique feature of Amazon’s EC2 T series instances is that they are burstable. T3 and T4 instances boost CPU performance when demand spikes; full CPU core power for one minute in Standard Mode and for as long as your workload requires in Unlimited Mode.

This is an advantage for applications requiring modest CPU usage, but occasionally needing to handle spikes in demand.

So, how do you choose between the t3, t3a, and t4g instances? Here’s what you need to know. We’ll also include a handy tool to help you pick the right type and size of instances for your requirements.

What Are Amazon EC2 T3 Instances?

An Amazon T3 instance is a General Purpose compute instance on the Amazon Web Services (AWS) platform. T3 instances use Intel Xeon Platinum 800 series processors from the first or second generation. With this processor line, you can expect all core Turbo CPU clock speeds of 3.1 GHz. And depending on the size of the instance, a baseline performance of between 5% and 40%.

T3 instances also work with the AWS Nitro System.

This is an AWS feature that takes advantage of a lightweight hypervisor and dedicated hardware to boost performance, availability, and security at a low cost (up to 30% price-performance over T2 instances).

Also, T3 instances accumulate credits when they perform below their baseline levels. A single CPU credit is equivalent to one vCPU operating at 100% load for one minute.

The credits system enables T-Series instances to run in Burstable mode at no extra cost to you.

Only when the credits have run out and the instance is yet to accumulate the balance after 24 hours, do extra charges kick in at $0.05 per vCPU hour (Standard mode). In Unlimited mode, you pay that additional $0.05 per vCPU-hour for operating in Burstable mode for prolonged periods.

Oh, another thing. Prices for T3 instances, and others in the T-Series, depend on factors such as your operating system, AWS Region, and payment plan, as well as your choice of instance type, size, vCPU, memory, and more. Check this out:

What Are Amazon EC2 T3a Instances?

Amazon T3a instances share similar features with T3 instances, except that they run on AMD EPYC 7000 series processors, clocked at 2.5 GHz (all core Turbo). With these and other features, AWS claims T3a instances can save you 10% over compatible T3.

What Are Amazon EC2 T4g Instances?

T4g instances feature the same credits system, AWS Nitro System, and Burstable mode as T3 instances. Also, T4g instances feature the latest, custom-built, Arm-based AWS Graviton2 processors. Compared to T3 instances, these AWS-specific processors can deliver up to 40% more performance for the same price.

T3 Vs. T3a Vs. T4g Instance Types: Side-By-Side Comparison Table

Here’s a quick table to help you compare T3, T3a, and T4g instances side-by-side so you can decide which is best for your computing needs.

	Amazon EC2 T3 instances	Amazon EC2 T3a instances	Amazon EC2 T4 instances
Speciality	Consistent CPU, memory, and network capacity, but can burst to handle peak loads	Cost-efficient processors Consistent CPU, memory, and network capacity, but can burst to handle peak loads	Most cost-efficient processors in the T-Series Consistent CPU, memory, and network capacity, but can burst to handle peak loads
Processor	3.1 GHz Intel Xeon Platinum 8000 Series (Skylake-SP or Cascade Lake)	2.5 GHz AMD EPYC 7000 Series	Arm AWS Graviton2
Max number of cores per instance	4	4	8
Max memory	32 GiB	32 GiB	32 GiB
Instance sizes available – baseline performance per vCPU – CPU credits earned per hour	t3.nano – 5% – 6 t3.micro – 10% – 12 t3.small – 20% – 24 t3.medium – 20% – 24 t3.large – 30% – 36 t3.xlarge – 40% – 96 t3.2xlarge – 40% – 192	t3a.nano – 5% – 6 t3a.micro – 10% – 12 t3a.small – 20% – 24 t3a.medium – 20% – 24 t3a.large – 30% – 36 t3a.xlarge – 40% – 96 t3a.2xlarge – 40% – 192	t4g.nano – 5% – 6 t4g.micro – 10% – 12 t4g.small – 20% – 24 t4g.medium – 20% – 24 t4g.large – 30% – 36 t4g.xlarge – 40% – 96 t4g.2xlarge – 40% – 192
Burstable Mode	Unlimited Mode enabled by default	Unlimited Mode enabled by default	Unlimited Mode enabled by default
Network bandwidth	Up to 5 Gbps	Up to 5 Gbps	Up to 5 Gbps
EBS bandwidth	Up to 2.085 Gbps	Up to 2.085 Gbps	Up to 2.085 Gbps, and 2.780 Gbps for largest three instance sizes
Best use cases	General purpose workloads with moderate CPU, memory, and network utilization, including small and mid-size virtual desktops, databases, and business-critical apps	General purpose workloads with moderate CPU, memory, and network utilization. Save 10% over T3 instance prices	General purpose workloads with moderate CPU, memory, and network utilization. Save up to 40% over T3 instance pricing

How To Choose The Right Instance Types Hassle-Free

If you’re an engineer, you want your applications to perform optimally even during peak demand. And if you are in finance or FinOps, you want to maximize ROI by minimizing costs.

Yet, because trial-and-error is often involved in configuring instances, reaching that price-performance goal can be time-consuming, expensive, and frustrating.

But what if you could achieve both?

Using CloudZero Advisor, you can get recommendations on the best instances for your application based on factors such as AWS service, pricing, AWS region, resource type, and more. Check out CloudZero Advisor here (it’s free).