Skip to main content

H200 vs L40S

The H200 delivers 5.5× the FP16 compute of the L40S while costing 4.0× more on average, making it 37% more cost-efficient per TFLOP.

Spec Comparison

Spec
H200
L40S
Architecture
Hopper
Ada Lovelace
VRAM
141 GB
48 GB
FP16 TFLOPS
1,979
362
Memory Bandwidth
4,800 GB/s
864 GB/s
MSRP
$40,000
$14,000
Used/Market Price
N/A
$9,000

Live Pricing

H200
Cheapest
$1.19/hr
Median
$12/hr
Average
$33/hr
Providers
9
Instances
133
L40S
Cheapest
$0.320/hr
Median
$3.00/hr
Average
$8.92/hr
Providers
19
Instances
180

Cost Efficiency

Metric
H200
L40S
$/TFLOP/hr
$0.0060
$0.0083
$/GB-VRAM/hr
$0.0849
$0.0626

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

Provider Availability

H200 Cloud Pricing

133 instances
ProviderGPUsVRAMRegionTypePrice/hr
CHEAPESTVerda1x141 GBEU-CentralSpot$1.19Rent
Nebius1x141 GBeu-north1Spot$1.45Rent
Nebius1x141 GBEU-WestSpot$1.45Rent
Nebius1x141 GBUS-CentralSpot$1.45Rent
RunPod1x141 GBUS-EastSpot$2.29Rent
RunPod1x141 GBEU-FR-1Spot$2.29Rent
RunPod1x141 GBEUR-IS-2Spot$2.29Rent
RunPod1x141 GBEUR-IS-4Spot$2.29Rent
RunPod1x141 GBUS-CA-2Spot$2.29Rent
RunPod1x141 GBUS-GA-2Spot$2.29Rent

L40S Cloud Pricing

180 instances
ProviderGPUsVRAMRegionTypePrice/hr
CHEAPESTVerda1x48 GBEU-CentralSpot$0.320Rent
AWS1x44.7 GBUS-EastSpot$0.446Rent
Crusoe1x48 GBUS-WestSpot$0.500Rent
AWS1x44.7 GBUS-EastSpot$0.532Rent
AWS1x44.7 GBAsia-PacificSpot$0.560Rent
AWS1x44.7 GBUS-EastSpot$0.587Rent
Verda2x96 GBEU-CentralSpot$0.640Rent
AWS1x44.7 GBAsia-PacificSpot$0.669Rent
AWS1x44.7 GBeu-north-1Spot$0.723Rent
AWS1x44.7 GBeu-north-1Spot$0.723Rent

Compare with Other GPUs