Skip to main content

A100 vs L40S

The A100 delivers 1.2× the FP16 compute of the A100 while costing 3.2× more on average, making the L40S 268% more cost-efficient per TFLOP.

Spec Comparison

Spec
A100
L40S
Architecture
Ampere
Ada Lovelace
VRAM
80 GB
48 GB
FP16 TFLOPS
312
362
Memory Bandwidth
2,000 GB/s
864 GB/s
MSRP
$10,000
$14,000
Used/Market Price
$7,000
$9,000

Live Pricing

A100
Cheapest
$0.0800/hr
Median
$8.52/hr
Average
$13/hr
Providers
14
Instances
358
L40S
Cheapest
$0.260/hr
Median
$2.69/hr
Average
$7.94/hr
Providers
20
Instances
213

Cost Efficiency

Metric
A100
L40S
$/TFLOP/hr
$0.0273
$0.0074
$/GB-VRAM/hr
$0.1065
$0.0560

Lower $/TFLOP and $/GB-VRAM is better. Based on median hourly pricing.

Provider Availability

A100 Cloud Pricing

358 instances
ProviderGPUsVRAMRegionTypePrice/hr
CHEAPESTVast.ai1x40 GBEU-CentralSpot$0.0800Rent
Vast.ai1x40 GBEU-CentralOn-Demand$0.0934Rent
Vultr1x4 GBewrOn-Demand$0.123Rent
Vultr1x4 GBfraOn-Demand$0.123Rent
Vultr1x4 GBsjcOn-Demand$0.123Rent
Vultr1x4 GBnrtOn-Demand$0.123Rent
Vultr1x8 GBewrOn-Demand$0.247Rent
Vultr1x8 GBnrtOn-Demand$0.247Rent
Vultr1x8 GBblrOn-Demand$0.247Rent
Verda1x40 GBEU-CentralSpot$0.252Rent

L40S Cloud Pricing

213 instances
ProviderGPUsVRAMRegionTypePrice/hr
CHEAPESTRunPod1x48 GBEU-NL-1Spot$0.260Rent
RunPod1x48 GBOC-AU-1Spot$0.260Rent
RunPod1x48 GBUS-TX-4Spot$0.260Rent
Verda1x48 GBEU-CentralSpot$0.320Rent
AWS1x44.7 GBUS-EastSpot$0.469Rent
Crusoe1x48 GBUS-WestSpot$0.500Rent
RunPod2x96 GBEU-NL-1Spot$0.520Rent
RunPod2x96 GBOC-AU-1Spot$0.520Rent
RunPod2x96 GBUS-TX-4Spot$0.520Rent
AWS1x44.7 GBUS-EastSpot$0.581Rent

Compare with Other GPUs