No FP8
The A100 is an Ampere GPU. NVIDIA lists FP16, BF16 and INT8 Tensor Core throughput for it but not FP8, which arrived with Hopper and Ada. Models quantized to FP8 run better on an H100, L40S or L4.
GPU rental prices · Ampere
NVIDIA A100 rental prices
NVIDIA A100 rental starts at $1.79 per GPU-hour (Runpod, PCIe 80 GB) across the 8 offers we track from 4 providers. The typical price is $2.19, and one GPU running all month costs about $1,307 at the lowest rate. Check the memory size before you compare prices. The A100 comes with 40 GB or 80 GB, and some 40 GB offers cost more per hour than 80 GB ones elsewhere.
- Lowest per GPU-hour
- $1.79 Runpod
- Typical per GPU-hour
- $2.19 median, USD
- One GPU all month
- $1,307 at the lowest rate
- Providers
- 4 8 offers
On-demand prices
NVIDIA A100 price per GPU-hour
Every A100 offer in our catalog, cheapest first. Multi-GPU machines are divided by their GPU count, so you can compare one GPU with one GPU. Monthly cost assumes the GPU runs all 730 hours.
| Provider and offer | Version | GPU memory | GPUs | Per GPU-hour | Per month, 1 GPU | Checked |
|---|---|---|---|---|---|---|
| RunpodRunpod Pod · NVIDIA A100 PCIe | PCIe 80 GB | 80 GB | 1 | $1.79 | $1,307 | 3 Nov 2026 |
| RunpodRunpod Pod · NVIDIA A100 SXM 80GB | SXM 80 GB | 80 GB | 1 | $1.79 | $1,307 | 3 Nov 2026 |
| LambdaLambda NVIDIA A100 PCIe 40GB · 1 GPU | PCIe 40 GB | 40 GB | 1 | $1.99 | $1,453 | 3 Nov 2026 |
| LambdaLambda NVIDIA A100 SXM 40GB · 1 GPU | SXM 40 GB | 40 GB | 1 | $1.99 | $1,453 | 3 Nov 2026 |
| VultrVultr NVIDIA A100 PCIe · 1 GPU | PCIe 80 GB | 80 GB | 1 | $2.40 | $1,750 | 3 Nov 2026 |
| VultrVultr NVIDIA A100 PCIe · 8 GPUs | PCIe 80 GB | 80 GB | 8$19.178/hr total | $2.40 | $1,750 | 3 Nov 2026 |
| PaperspacePaperspace A100 · 1 GPU | 40 GB | 40 GB | 1 | $3.09 | $2,256 | 3 Nov 2026 |
| PaperspacePaperspace A100-80G · 1 GPU | SXM 80 GB | 80 GB | 1 | $3.18 | $2,321 | 3 Nov 2026 |
On-demand list prices before tax, from each provider's pricing page. Community and reserved prices, where a provider offers them, are lower; see the notes below. Full offer details, storage and billing terms.
Before you rent
What changes the price
40 GB can cost more
Lambda's 40 GB A100s cost $1.99 an hour and Paperspace's $3.09, while Runpod rents an 80 GB A100 for $1.79.
Community prices
Runpod's Community Cloud lists the A100 PCIe at $1.19 and the A100 SXM at $1.39 an hour.
AI models
Which models fit on the A100
Memory to load each model size on 80 GB GPUs, with a 20% allowance for runtime buffers, and what that many GPUs cost per hour at the lowest rate above. Long contexts and many parallel users need more memory. More than one GPU only works if your inference server splits the model across them.
| Model size | 4-bit | 8-bit | 16-bit |
|---|---|---|---|
| 8BLlama 3.1 8B, Qwen3 8B | 4.5 GB1 GPU · $1.79/hr | 9.0 GB1 GPU · $1.79/hr | 17.9 GB1 GPU · $1.79/hr |
| 32BQwen3 32B, Gemma 4 31B | 17.9 GB1 GPU · $1.79/hr | 35.8 GB1 GPU · $1.79/hr | 71.6 GB1 GPU · $1.79/hr |
| 70BLlama 3.3 70B, Qwen 72B | 39.2 GB1 GPU · $1.79/hr | 78.3 GB1 GPU · $1.79/hr | 156.5 GB2 GPUs · $3.58/hr |
| 120Bgpt-oss-120b class | 67.1 GB1 GPU · $1.79/hr | 134.2 GB2 GPUs · $3.58/hr | 268.3 GB4 GPUs · $7.16/hr |
| 685BDeepSeek-V3.2 | 382.8 GB5 GPUs · $8.95/hr | 765.6 GB10 GPUs · more than one 8-GPU server | 1,531.1 GB20 GPUs · more than one 8-GPU server |
Weights only, from the parameter count. Check a specific model, context length and user count in the model hardware calculator.
Hardware
NVIDIA A100 specifications
| Specification | A100 80GB PCIe | A100 80GB SXM |
|---|---|---|
| GPU memory | 80 GB HBM2e | 80 GB HBM2e |
| Memory bandwidth | 1,935 GB/s | 2,039 GB/s |
| FP16 Tensor | 312 TFLOPS (624 sparse) | 312 TFLOPS (624 sparse) |
| Max power | 300 W | 400 W |
| NVLink | Bridge for 2 GPUs, 600 GB/s | 600 GB/s |
From NVIDIA's product pages. Tensor figures marked “sparse” assume structured sparsity; dense throughput is half.
Questions
A100 rental FAQ
How much does it cost to rent an A100?
$1.79 per GPU-hour is the lowest on-demand price we track, from Runpod. The typical price is $2.19, and one GPU running all month costs about $1,307 at the lowest rate.
Can an A100 run a 70B model?
On one 80 GB A100, a 70B model fits comfortably only at 4-bit: the weights need about 39.2 GB with a 20% allowance. At 8-bit they take 78.3 GB, which leaves under 2 GB for the conversation cache, so plan on two GPUs. A 40 GB A100 needs twice as many cards.
Should I rent an A100 or an H100?
The H100 SXM has about 1.6 times the memory bandwidth of the A100 SXM (3.35 TB/s against 2,039 GB/s) and adds FP8. At Runpod the H100 SXM costs $3.99 an hour and the A100 SXM $1.79, so the A100 is still the cheaper choice when a model fits and speed is not critical.
Compare
Other GPU rental prices
Prices and availability change, and we don't check stock. Confirm the current price and region with the provider before you rent. Sources: www.nvidia.com, www.runpod.io, lambda.ai, www.vultr.com, docs.digitalocean.com . See our affiliate disclosure.