Compute Comparison
vs
All comparisons →

CDNA 4

AMD Instinct MI355X

wins

4

The AMD Instinct MI355X is AMD's CDNA 4 flagship.

Large LLM inferenceAgent workloadsMemory-bound AI

CDNA 3

AMD Instinct MI300X

wins

0

The AMD MI300X leads all GPUs on memory capacity (192GB HBM3) and bandwidth (5.3 TB/s), making it exceptional for serving very large language models that don't fit on H100.

Memory-bound LLM inferenceLarge model servingHPC

Performance comparison

Visual bars — winner highlighted

AMD Instinct MI355X

Metric

AMD Instinct MI300X

FP16 TFLOPS

2.6k T
1.3k T

Bandwidth

8.0k GB/s
5.3k GB/s

VRAM

2.9k GB
1.9k GB

TDP (lower=better)

750 W
750 W

TFLOPS/watt

3.49
1.74

GB/s per watt

10.67
7.07

Full specification table

SpecAMD Instinct MI355XAMD Instinct MI300X
ArchitectureCDNA 4CDNA 3
VRAM288 GB HBM3e192 GB HBM3
VRAM typeHBM3eHBM3
Memory bandwidth8,000 GB/s5,300 GB/s
FP16 TFLOPS2,615 TFLOPS1,307 TFLOPS
TDP750 W750 W
TFLOPS/watt3.49 T/W1.74 T/W
GB/s per watt10.67 GB/s/W7.07 GB/s/W
NVLinkNoNo
Release year20252023

Live cloud pricing

On-demand $/hr across providers — updated in real time

Loading prices…

Power efficiency analysis

TFLOPS/watt and GB/s/watt — critical for data center TCO

AMD Instinct MI355X

FP16 TFLOPS2,615 TFLOPS
TDP750 W
TFLOPS/watt3.487 T/W
GB/s per watt10.67 GB/s/W

AMD Instinct MI300X

FP16 TFLOPS1,307 TFLOPS
TDP750 W
TFLOPS/watt1.743 T/W
GB/s per watt7.07 GB/s/W

TFLOPS/watt measures compute efficiency — how much AI throughput you get per watt of power consumed. For data centers with PUE of 1.2–1.5, a 10% improvement in TFLOPS/watt translates directly to lower electricity costs and cooling requirements. GB/s/watt measures memory bandwidth efficiency, which is the binding constraint for memory-bound LLM inference workloads.

When to choose each GPU

Choose AMD Instinct MI355X for:

  • Large LLM inference
  • Agent workloads
  • Memory-bound AI

Choose AMD Instinct MI300X for:

  • Memory-bound LLM inference
  • Large model serving
  • HPC