| NVIDIA RTX 3090 |
24 GB |
Salad |
$0.12 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 936 GB/s
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Salad
|
| NVIDIA RTX 3090 Ti |
24 GB |
Salad |
$0.15 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 1,008 GB/s
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Salad
|
| NVIDIA RTX 4080 |
16 GB |
Salad |
$0.15 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 717 GB/s
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Salad
|
| NVIDIA RTX 3090 |
24 GB |
TensorDock |
$0.20 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 936 GB/s
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit TensorDock
|
| NVIDIA RTX 5080 |
16 GB |
Salad |
$0.22 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Blackwell
- Memory bandwidth:
- n/a
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Salad
|
| NVIDIA RTX 5090 |
32 GB |
Salad |
$0.29 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Blackwell
- Memory bandwidth:
- n/a
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Salad
|
| NVIDIA RTX 4090 |
24 GB |
Salad |
$0.30 |
$0.18 |
Variable |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 1,008 GB/s
- FP16 manufacturer spec:
- 330 TFLOPS
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Salad
|
| NVIDIA RTX 4090 |
24 GB |
TensorDock |
$0.35 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 1,008 GB/s
- FP16 manufacturer spec:
- 330 TFLOPS
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit TensorDock
|
| NVIDIA RTX 4090 |
24 GB |
Vast.ai |
$0.39 |
$0.29 |
Variable |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 1,008 GB/s
- FP16 manufacturer spec:
- 330 TFLOPS
- Tier:
- Consumer
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Vast.ai
|
| NVIDIA RTX 4090 |
24 GB |
RunPod |
$0.44 |
$0.34 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 1,008 GB/s
- FP16 manufacturer spec:
- 330 TFLOPS
- Tier:
- Consumer
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA A40 |
48 GB |
RunPod |
$0.44 |
$0.35 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 696 GB/s
- FP16 manufacturer spec:
- 150 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA L4 |
24 GB |
RunPod |
$0.44 |
$0.39 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 300 GB/s
- FP16 manufacturer spec:
- 242 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA RTX 3090 |
24 GB |
RunPod |
$0.46 |
$0.22 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 936 GB/s
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA RTX 5090 |
32 GB |
RunPod |
$0.69 |
$0.49 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Blackwell
- Memory bandwidth:
- n/a
- FP16 manufacturer spec:
- n/a
- Tier:
- Consumer
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA RTX 6000 Ada |
48 GB |
TensorDock |
$0.75 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 960 GB/s
- FP16 manufacturer spec:
- 729 TFLOPS
- Tier:
- Data center
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit TensorDock
|
| NVIDIA RTX 6000 Ada |
48 GB |
RunPod |
$0.77 |
$0.74 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 960 GB/s
- FP16 manufacturer spec:
- 729 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA L40 |
48 GB |
RunPod |
$0.82 |
$0.69 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 864 GB/s
- FP16 manufacturer spec:
- 362 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA L40 |
48 GB |
TensorDock |
$0.95 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 864 GB/s
- FP16 manufacturer spec:
- 362 TFLOPS
- Tier:
- Data center
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit TensorDock
|
| NVIDIA L40S |
48 GB |
RunPod |
$0.99 |
$0.79 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 864 GB/s
- FP16 manufacturer spec:
- 362 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA L40 |
48 GB |
CoreWeave |
$1.25 |
$0.78 |
High |
US, EU |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 864 GB/s
- FP16 manufacturer spec:
- 362 TFLOPS
- Tier:
- Data center
- Regions:
- US, EU
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit CoreWeave
|
| NVIDIA A10 |
24 GB |
Lambda |
$1.29 |
n/a |
Medium |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 600 GB/s
- FP16 manufacturer spec:
- 125 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Lambda
|
| NVIDIA A100 80GB PCIe |
80 GB |
RunPod |
$1.49 |
$1.19 |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 2,000 GB/s
- FP16 manufacturer spec:
- 624 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA A100 80GB SXM |
80 GB |
RunPod |
$1.49 |
n/a |
Variable |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 1,935 GB/s
- FP16 manufacturer spec:
- 624 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit RunPod
|
| NVIDIA A100 80GB PCIe |
80 GB |
TensorDock |
$1.50 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 2,000 GB/s
- FP16 manufacturer spec:
- 624 TFLOPS
- Tier:
- Data center
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit TensorDock
|
| NVIDIA A100 80GB SXM |
80 GB |
TensorDock |
$1.80 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 1,935 GB/s
- FP16 manufacturer spec:
- 624 TFLOPS
- Tier:
- Data center
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit TensorDock
|
| NVIDIA A100 40GB SXM |
40 GB |
Lambda |
$1.99 |
n/a |
Medium |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 1,550 GB/s
- FP16 manufacturer spec:
- 624 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Lambda
|
| NVIDIA L40S |
48 GB |
CoreWeave |
$2.25 |
$0.99 |
High |
US, EU |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ada Lovelace
- Memory bandwidth:
- 864 GB/s
- FP16 manufacturer spec:
- 362 TFLOPS
- Tier:
- Data center
- Regions:
- US, EU
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit CoreWeave
|
| NVIDIA H100 SXM |
80 GB |
TensorDock |
$2.25 |
n/a |
High |
Global / varies |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Hopper
- Memory bandwidth:
- 3,350 GB/s
- FP16 manufacturer spec:
- 1,979 TFLOPS
- Tier:
- Data center
- Regions:
- Global / varies
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit TensorDock
|
| NVIDIA A100 80GB SXM |
80 GB |
CoreWeave |
$2.70 |
$1.21 |
High |
US, EU |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 1,935 GB/s
- FP16 manufacturer spec:
- 624 TFLOPS
- Tier:
- Data center
- Regions:
- US, EU
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit CoreWeave
|
| NVIDIA A100 80GB SXM |
80 GB |
Lambda |
$2.79 |
n/a |
Medium |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Ampere
- Memory bandwidth:
- 1,935 GB/s
- FP16 manufacturer spec:
- 624 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Lambda
|
| NVIDIA H100 PCIe |
80 GB |
Lambda |
$3.29 |
n/a |
Medium |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Hopper
- Memory bandwidth:
- 2,000 GB/s
- FP16 manufacturer spec:
- 1,513 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Lambda
|
| AMD Instinct MI300X |
192 GB |
CoreWeave |
$3.49 |
n/a |
Low |
US |
|
Details- Vendor:
- AMD
- Architecture:
- CDNA 3
- Memory bandwidth:
- 5,300 GB/s
- FP16 manufacturer spec:
- 1,307 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit CoreWeave
|
| NVIDIA H200 |
141 GB |
Lambda |
$3.49 |
n/a |
Low |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Hopper
- Memory bandwidth:
- 4,800 GB/s
- FP16 manufacturer spec:
- 1,979 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Lambda
|
| NVIDIA H100 SXM |
80 GB |
Lambda |
$3.99 |
n/a |
Medium |
US |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Hopper
- Memory bandwidth:
- 3,350 GB/s
- FP16 manufacturer spec:
- 1,979 TFLOPS
- Tier:
- Data center
- Regions:
- US
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit Lambda
|
| NVIDIA H100 SXM |
80 GB |
CoreWeave |
$6.16 |
$2.46 |
High |
US, EU |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Hopper
- Memory bandwidth:
- 3,350 GB/s
- FP16 manufacturer spec:
- 1,979 TFLOPS
- Tier:
- Data center
- Regions:
- US, EU
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit CoreWeave
|
| NVIDIA H200 |
141 GB |
CoreWeave |
$6.31 |
$2.62 |
High |
US, EU |
|
Details- Vendor:
- NVIDIA
- Architecture:
- Hopper
- Memory bandwidth:
- 4,800 GB/s
- FP16 manufacturer spec:
- 1,979 TFLOPS
- Tier:
- Data center
- Regions:
- US, EU
- Last verified:
- Source:
- View pricing source
- Provider:
- Visit CoreWeave
|
