A100
NVIDIA's flagship data center GPU with 80GB of HBM2e memory. Designed for AI training and inference at scale.
Average Price: $0.002203/s ($7.9302/h)
Available from 6 vendors
Serverless GPU Prices
Provider | GPU RAM | RAM | vCPUs | Price |
|---|---|---|---|---|
| Inferless | 12 | 31 | 3 | 0.000233/s 0.8388/h |
| Fal.ai | 40 | 0.0003/s 1.08/h | ||
| Fal.ai | 40 | 0.0003/s 1.08/h | ||
| Modal | 40 | 0.000583/s 2.0988/h | ||
| Modal | 40 | 0.000583/s 2.0988/h | ||
| Modal | 80 | 0.000694/s 2.4984/h | ||
| Modal | 80 | 0.000694/s 2.4984/h | ||
| Inferless | 40 | 100 | 10 | 0.000745/s 2.682/h |
| Runpod | 80 | 0.00076/s 2.736/h | ||
| Runpod | 80 | 0.00076/s 2.736/h | ||
| Beam.cloud | 40 | 0.000764/s 2.7504/h | ||
| Beam.cloud | 80 | 0.000955/s 3.438/h | ||
| Beam.cloud | 40 | 0.000972/s 3.4992/h | ||
| Replicate | 40 | 72 | 10 | 0.00115/s 4.14/h |
| Runpod | 80 | 0.0013/s 4.68/h | ||
| Replicate | 80 | 144 | 10 | 0.0014/s 5.04/h |
| Replicate | 80 | 144 | 10 | 0.0014/s 5.04/h |
| Replicate | 80 | 144 | 10 | 0.0014/s 5.04/h |
| Inferless | 80 | 200 | 20 | 0.001491/s 5.3676/h |
| Inferless | 8 | 21 | 2 | 0.00159/s 5.724/h |
| Replicate | 160 | 288 | 20 | 0.0028/s 10.08/h |
| Replicate | 160 | 288 | 20 | 0.0028/s 10.08/h |
| Replicate | 320 | 576 | 40 | 0.0056/s 20.16/h |
| Replicate | 320 | 576 | 40 | 0.0056/s 20.16/h |
| Replicate | 640 | 960 | 80 | 0.0112/s 40.32/h |
| Replicate | 640 | 960 | 80 | 0.0112/s 40.32/h |