AWS Inferentia2 vs AWS Trainium3

The Trainium3 delivers 13.2x the FP8 throughput of the Inferentia2 (2,517 vs 190 TFLOPS dense). The Trainium3 also carries 112 GB more memory (144 GB vs 32 GB).

Inferentia2: NeuronCore-v2, 2022 Trainium3: NeuronCore-v4, 2025

FP8 dense lead

13.2x

Trainium3: 2,517 vs 190 TFLOPS

Memory

32 vs 144 GB

112 GB more for the Trainium3

Bandwidth

820 GB/s vs 4.9 TB/s

Trainium3 moves data faster

Download comparison sheet Printable version →

AWS Inferentia2

Full specs →

What is the AWS Inferentia2?

The Inferentia2 is an ASIC from AWS on the NeuronCore-v2 architecture, an accelerator built for AI workloads. Inferentia2 full specifications →

Can I rent the Inferentia2 in the cloud?

Flopper is not tracking any live rental listings for the Inferentia2 yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Inferentia2 compare with other accelerators?

This page sets the Inferentia2 against the Trainium3. The Inferentia2 is the older generation of the two. Both are AWS parts. Flopper also compares it head to head with the L4 24GB, across every published precision, memory and power. Inferentia2 vs L4 24GB →

Where do the Inferentia2 figures come from?

Every Inferentia2 figure on this page traces to AWS's own datasheet or specification table. How Flopper verifies specifications →

AWS Trainium3

Leads on FP8 Full specs →

What is the AWS Trainium3?

The Trainium3 is an ASIC from AWS on the NeuronCore-v4 architecture, an accelerator built for AI workloads. Trainium3 full specifications →

Can I rent the Trainium3 in the cloud?

Flopper is not tracking any live rental listings for the Trainium3 yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Trainium3 compare with other accelerators?

This page sets the Trainium3 against the Inferentia2. The Trainium3 is the newer generation of the two. Both are AWS parts. The compare tool builds the same comparison for any two parts in the catalogue. All GPU comparisons →

Where do the Trainium3 figures come from?

Every Trainium3 figure on this page traces to AWS's own datasheet or specification table. How Flopper verifies specifications →

Specifications

Inferentia2

Trainium3

Architecture

NeuronCore-v2NeuronCore-v4

Launch Year

20222025

Form Factor

——

Memory

32 GB144 GB

Memory Bandwidth

820 GB/s4.9 TB/s

TDP

——

Process Node

—3nm

Spec

Inferentia2

Trainium3

Architecture NeuronCore-v2NeuronCore-v4

Launch Year 20222025

Form Factor ——

Memory 32 GB144 GB

Memory Bandwidth 820 GB/s4.9 TB/s

TDP ——

Process Node —3nm

Performance (TFLOPS)

Inferentia2

Trainium3

48 TFLOPS 183 TFLOPS

190 TFLOPS 671 TFLOPS

2,517 TFLOPS sparse

190 TFLOPS 671 TFLOPS

2,517 TFLOPS sparse

190 TFLOPS 671 TFLOPS

2,517 TFLOPS sparse

190 TFLOPS 2,517 TFLOPS

sparse not published

FP6

2,517 TFLOPS

sparse not published

380 TOPS

Precision

Inferentia2

Trainium3

48 TFLOPS 183 TFLOPS

190 TFLOPS 671 TFLOPS

2,517 TFLOPS sparse

190 TFLOPS 671 TFLOPS

2,517 TFLOPS sparse

190 TFLOPS 671 TFLOPS

2,517 TFLOPS sparse

190 TFLOPS 2,517 TFLOPS

sparse not published

FP6

2,517 TFLOPS

sparse not published

380 TOPS

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

The Trainium3 is the newer part: NeuronCore-v4 (2025) against the Inferentia2's NeuronCore-v2 (2022).

The Trainium3 carries 112 GB more memory (144 GB against 32 GB), so larger models fit on one GPU before you shard. Bandwidth is 820 GB/s against 4.9 TB/s, which sets the ceiling on decode speed.

Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP32, TF32, BF16, FP16, FP8, FP4, INT8, Sparsity and TFLOPS.

Frequently asked questions

What is better, the Inferentia2 or the Trainium3?

The Trainium3, on throughput. At FP8 precision the Trainium3 reaches 2,517 TFLOPS dense against 190 TFLOPS for the Inferentia2. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the Inferentia2 or the Trainium3?

Choose the Trainium3 for training and high-throughput inference: it has 13.2x the FP8 throughput of the Inferentia2. It also carries 112 GB more memory, so larger models fit on one GPU before you shard.

Which has more memory, the Inferentia2 or the Trainium3?

The Trainium3 carries 144 GB of memory versus 32 GB for the Inferentia2. Memory bandwidth is 820 GB/s for the Inferentia2 and 4.9 TB/s for the Trainium3.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

Open in Full Compare Tool → Inferentia2 full specs Trainium3 full specs Estimate cluster cost in the calculator →

AWS Inferentia2 vs AWS Trainium3

AWS Inferentia2

What is the AWS Inferentia2?

Can I rent the Inferentia2 in the cloud?

How does the Inferentia2 compare with other accelerators?

Where do the Inferentia2 figures come from?

AWS Trainium3

What is the AWS Trainium3?

Can I rent the Trainium3 in the cloud?

How does the Trainium3 compare with other accelerators?

Where do the Trainium3 figures come from?

Specifications

Performance (TFLOPS)

FLOPS by Precision

What actually differs

Frequently asked questions

What is better, the Inferentia2 or the Trainium3?

Which should you choose, the Inferentia2 or the Trainium3?

Which has more memory, the Inferentia2 or the Trainium3?

Related comparisons

Get Comparison Updates

Stay Updated on GPU Specs