RTX4090 vs T4
The RTX4090 delivers 2.5× the FP16 compute of the T4 while costing 1.3× less on average, making it 221% more cost-efficient per TFLOP.
Spec Comparison
Spec
RTX4090
T4
Architecture
Ada Lovelace
Older
VRAM
24 GB
16 GB
FP16 TFLOPS
165
65
Memory Bandwidth
1,008 GB/s
320 GB/s
MSRP
$1,599
N/A
Used/Market Price
$1,200
$1,500
Live Pricing
Cost Efficiency
Metric
RTX4090
T4
$/TFLOP/hr
$0.0044
$0.0140
$/GB-VRAM/hr
$0.0299
$0.0568
Lower $/TFLOP and $/GB-VRAM is better. Based on median hourly pricing.
Provider Availability
RTX4090 Cloud Pricing
96 instances| Provider | GPUs | VRAM | Region | Type | Price/hr | |
|---|---|---|---|---|---|---|
| CHEAPESTVast.ai | 1x | 24 GB | Spot | $0.131 | Rent | |
| RunPod | 1x | 24 GB | CA | Spot | $0.200 | Rent |
| RunPod | 1x | 24 GB | FR | Spot | $0.200 | Rent |
| RunPod | 1x | 24 GB | SE | Spot | $0.200 | Rent |
| RunPod | 1x | 24 GB | SK | Spot | $0.200 | Rent |
| RunPod | 1x | 24 GB | TT | Spot | $0.200 | Rent |
| RunPod | 1x | 24 GB | US | Spot | $0.200 | Rent |
| RunPod | 1x | 24 GB | EU-RO-1 | Spot | $0.290 | Rent |
| RunPod | 1x | 24 GB | EUR-IS-2 | Spot | $0.290 | Rent |
| RunPod | 1x | 24 GB | US-CA-2 | Spot | $0.290 | Rent |
T4 Cloud Pricing
1557 instances| Provider | GPUs | VRAM | Region | Type | Price/hr | |
|---|---|---|---|---|---|---|
| CHEAPESTAzure | 1x | 16 GB | Asia-Pacific | Spot | $0.0684 | Rent |
| Azure | 1x | 16 GB | US-Central | Spot | $0.0686 | Rent |
| Azure | 1x | 16 GB | Asia-Pacific | Spot | $0.0977 | Rent |
| Azure | 1x | 16 GB | US-Central | Spot | $0.0980 | Rent |
| GCP | 1x | 16 GB | me-west1-b | Spot | $0.112 | Rent |
| GCP | 1x | 16 GB | me-west1-c | Spot | $0.112 | Rent |
| GCP | 1x | 16 GB | EU-West | Spot | $0.117 | Rent |
| GCP | 1x | 16 GB | me-west1-b | Spot | $0.118 | Rent |
| GCP | 1x | 16 GB | me-west1-c | Spot | $0.118 | Rent |
| AWS | 1x | 16 GB | South America | Spot | $0.123 | Rent |