Compute Comparison
vs
All comparisons →

Ampere

NVIDIA RTX 3080 Ti 12GB

wins

2

The RTX 3080 Ti is a near-flagship Ampere consumer GPU with 12 GB GDDR6X and 68.2 TFLOPS FP16.

Budget inferenceSmall model fine-tuningExperimentation

Blackwell

NVIDIA RTX 5090 32GB

wins

2

The RTX 5090 is NVIDIA's Blackwell consumer flagship with 32 GB GDDR7 and 838 TFLOPS FP16.

Consumer AI inferenceFine-tuning 13B–30B modelsCost-effective Blackwell

Performance comparison

Visual bars — winner highlighted

NVIDIA RTX 3080 Ti 12GB

Metric

NVIDIA RTX 5090 32GB

FP16 Tensor TFLOPS (dense)

68 T
0 T

Bandwidth

912 GB/s
1.8k GB/s

VRAM

12 GB
32 GB

TDP (lower=better)

350 W
575 W

TFLOPS/watt

0.19
0.00

GB/s per watt

2.61
3.12

Full specification table

SpecNVIDIA RTX 3080 Ti 12GBNVIDIA RTX 5090 32GB
ArchitectureAmpereBlackwell
VRAM12 GB GDDR6X32 GB GDDR7
VRAM typeGDDR6XGDDR7
Memory bandwidth912 GB/s1,792 GB/s
FP16 Tensor TFLOPS (dense)68.2 TFLOPSn/a
TDP350 W575 W
TFLOPS/watt0.19 T/W0.00 T/W
GB/s per watt2.61 GB/s/W3.12 GB/s/W
NVLinkNoNo
Release year20212025

Live cloud pricing

On-demand pricing across providers — updated in real time

Per hourPer GB VRAM/hr
Loading prices…

Power efficiency analysis

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

NVIDIA RTX 3080 Ti 12GB

FP16 TFLOPS68.2 TFLOPS
TDP350 W
TFLOPS/watt0.195 T/W
GB/s per watt2.61 GB/s/W

NVIDIA RTX 5090 32GB

FP16 TFLOPSn/a
TDP575 W
TFLOPS/watt0.000 T/W
GB/s per watt3.12 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 NVIDIA RTX 3080 Ti 12GB for:

  • Budget inference
  • Small model fine-tuning
  • Experimentation

Weaknesses / Limitations

  • Only 12 GB VRAM — limits to very small quantized models
  • No BF16 hardware acceleration
  • Consumer-grade reliability — no ECC memory

Choose NVIDIA RTX 5090 32GB for:

  • Consumer AI inference
  • Fine-tuning 13B–30B models
  • Cost-effective Blackwell

Weaknesses / Limitations

  • 32 GB VRAM — cannot run 70B+ models unquantized
  • No NVLink — cannot pool VRAM across cards
  • 575W TDP — high power draw for a consumer card