NVIDIA A100 SXM4 rental pricing
Ampere architecture · 80 GB memory · 2.0 TB/s bandwidth · 400 W board power · Released 2020
Price history by provider
Every provider's cheapest daily on-demand rate for the NVIDIA A100 SXM4, last 30 days
One line per provider, reconstructed from recorded price changes for listings currently live. Click a name in the key to hide or show it.
At a glance
prices checked 5h ago- Cheapest in stock
- $0.55/hr
- Community · 1× A100 SXM4 · Vast.ai
- Median price
- $1.70/hr
- On-demand · 22 providers
- 90-day trend
- Flat
- Weekly median on-demand, 4 weeks
- Coverage
- 25 providers
- 56 price points
- $ per FP16 PFLOP-hour
- $4.36
- Rank 5 of 17 priced GPUs · lower is better
Why the SXM4 costs what it does
Against the other priced NVIDIA A100 variants: the price floor, what a GB-hour and a PFLOP-hour cost, and the specs that explain the gap.
The SXM4 rents from $0.55 per GPU-hour, 9% less than the PCIe at $0.61. Only 1 listing tracks the PCIe, so its floor is one provider's choice rather than a market rate.
The SXM4 is an SXM module, rated at 400 W against 250 W for the PCIe card. Hosting a module with that power and interconnect budget costs more than hosting a card, and the hourly rate carries it.
The SXM4 carries 80 GB against 40 GB, 100% more memory and it still rents for less; per GB-hour that is $0.007 against $0.015.
Memory bandwidth is 2.0 TB/s on the SXM4 against 1.6 TB/s on the PCIe, 31% more, which is what sets decode throughput for inference.
25 providers list the SXM4 against 1 for the PCIe; a deeper market tends to price lower.
Floors are the cheapest live listing of any pricing type; medians are of each provider's cheapest on-demand rate. Per GB-hour divides the floor by memory, per PFLOP-hour by dense throughput on the family's basis.
Every provider we track
Cheapest listing per provider and pricing type. Sold-out rows stay so you can see who normally carries it.
| Provider | Configuration | Price/GPU-hr | Checked | |
|---|---|---|---|---|
| 1× A100 SXM4 Community | $0.55 | 5h ago | View → | |
 Oblivus | 1× A100 SXM4 Reserved | $1.05 | 5h ago | View → |
 Thunder Compute | 1× A100 SXM4 | $1.09 | Stale · 42d ago | View → |
| 1× A100 SXM4 Spot | $1.16 | 5h ago | View → | |
 Spheron | 1× A100 SXM4 Spot | $1.19 | 5h ago | View → |
| 8× A100 SXM4 Spot | $1.21 | 5h ago | View → | |
 GPU.ai | 1× A100 SXM4 Community | $1.25 | 5h ago | View → |
 Beam | 1× A100 SXM4 | $1.36 | 9d ago | View → |
| GPU.ai | 1× A100 SXM4 | $1.38 | 5h ago | View → |
| 1× A100 SXM4 Community | $1.39 | 5h ago | View → | |
 Civo | 1× A100 SXM4 Reserved | $1.39 | 5h ago | View → |
 Shadeform Denvr Dataworks capacity | 8× A100 SXM4 | $1.40 | 5h ago | View → |
| Spheron | 1× A100 SXM4 | $1.48 | 5h ago | View → |
| 1× A100 SXM4 | $1.49 | Stale · 34d ago | View → | |
 Gcore | 1× A100 SXM4 | $1.53 | 5h ago | View → |
| Oblivus | 1× A100 SXM4 | $1.57 | 5h ago | View → |
 Lyceum | 1× A100 SXM4 | $1.59 | Stale · 36d ago | View → |
| 1× A100 SXM4 | $1.60 | 5h ago | View → | |
 Koyeb | 1× A100 SXM4 | $1.60 | 5h ago | View → |
| 1× A100 SXM4 | $1.79 | 5h ago | View → | |
| Civo | 1× A100 SXM4 | $1.79 | 5h ago | View → |
| 1× A100 SXM4 | $1.99 | 5h ago | View → | |
 Cerebrium | 1× A100 SXM4 | $2.10 | 5h ago | View → |
| 1× A100 SXM4 | $2.32 | 5h ago | View → | |
| 1× A100 SXM4 | $2.50 | Stale · 41d ago | View → | |
 Cirrascale | 8× A100 SXM4 Reserved | $2.60 | 6d ago | View → |
| 8× A100 SXM4 | $2.70 | 5h ago | View → | |
| 8× A100 SXM4 | $3.43 | 5h ago | View → | |
| 1× A100 SXM4 | $4.00 | 8d ago | View → | |
| 8× A100 SXM4 | $4.10 | 11h ago | View → | |
 Replicate | 1× A100 SXM4 | $5.04 | 5h ago | View → |
| Replicate | 4× A100 SXM4 Reserved | $5.04 | 5h ago | View → |
| 1× A100 SXM4 | $1.81 | 5h ago | View → |
Per GPU-hour. Spot, reserved and community pricing differ from on-demand; click through for the provider's current figure.
What can the NVIDIA A100 SXM4 run?
Memory for the weights plus one sequence of context, how many GPUs hold it, and what that replica costs at the cheapest in-stock on-demand rate.
| Model | Params | Memory | GPUs | $/hr |
|---|---|---|---|---|
| gpt-oss-20b OpenAI | 20.9B· 3.6B active | 20 GiB | 1 | $1.09 |
| Gemma 3 27B Google | 27.4B | 26 GiB | 1 | $1.09 |
| Qwen3-32B Alibaba | 32.8B | 32 GiB | 1 | $1.09 |
| Llama 3.3 70B Instruct Meta | 70.6B | 67 GiB | 1 | $1.09 |
| gpt-oss-120b OpenAI | 117B· 5.1B active | 109 GiB | 2 | $2.18 |
| Qwen3-235B-A22B Alibaba | 235B· 22.0B active | 220 GiB | 4 | $4.36 |
| DeepSeek-R1 DeepSeek | 684B· 37.0B active | 638 GiB | 16 | $17.44 |
| Kimi K3 Moonshot AI | 2.78T· 104B active | 2.53 TiB | Over 16 | — |
Sized for one sequence at 4K context with 16-bit KV cache and 10% of memory held back. Serving many users at once needs more; the LLM GPU calculator sizes for concurrency and throughput.
NVIDIA A100 SXM4 pricing questions
How much does it cost to rent a NVIDIA A100 SXM4 per hour?
The cheapest NVIDIA A100 SXM4 in stock right now is $0.55 per GPU-hour community from Vast.ai for the 1× A100 SXM4 configuration. The median on-demand rate across 22 providers is $1.70 per GPU-hour.
How much does a NVIDIA A100 SXM4 cost per month?
Running one NVIDIA A100 SXM4 around the clock for a month (730 hours) costs about $402 at the cheapest in-stock rate of $0.55 per hour. At the median on-demand rate it is about $1,237 a month. Reserved terms usually come in below that; ask the provider for a quote.
Which provider has the cheapest NVIDIA A100 SXM4?
Vast.ai has the cheapest NVIDIA A100 SXM4 in stock right now at $0.55 per GPU-hour community for the 1× A100 SXM4 configuration. That is a community rate; see the provider table for the cheapest on-demand listing. Rates change often, so check the provider table for today's figure before you commit.
Are NVIDIA A100 SXM4 rental prices going up or down?
The weekly median on-demand rate for the NVIDIA A100 SXM4 has been flat, moving 0.0% over the last 4 weeks, from $1.79 to $1.79 per GPU-hour.
What is the difference between on-demand, spot and reserved NVIDIA A100 SXM4 pricing?
On-demand is pay-as-you-go with no commitment. Spot is spare capacity sold at a discount that the provider can reclaim at short notice. Reserved is a committed term, usually months, in exchange for a lower rate. Community is capacity from individual hosts on a marketplace, cheaper and less uniform. This week's median rates for the NVIDIA A100 SXM4 are on-demand $1.79 (23 providers), spot $1.19 (3 providers), reserved $2.00 (4 providers), community $1.11 (3 providers), per GPU-hour.
How many cloud providers rent the NVIDIA A100 SXM4?
We track 25 providers with 56 live prices for the NVIDIA A100 SXM4. Sold-out listings stay in the table so you can see who normally carries it, but never set the cheapest figure.
What LLMs can you run on one NVIDIA A100 SXM4?
With 8-bit weights and a single sequence of context, one NVIDIA A100 SXM4 holds gpt-oss-20b (20.9B), Gemma 3 27B (27.4B), Qwen3-32B (32.8B), Llama 3.3 70B Instruct (70.6B). See the model fit table for the full ladder.
How does the NVIDIA A100 SXM4 compare on price per PFLOP-hour?
At its cheapest in-stock on-demand rate the NVIDIA A100 SXM4 delivers dense FP16 compute at $4.36 per PFLOP-hour, which ranks 5 of 17 priced GPUs we cover on that basis. Lower is better: it is the rental price divided by peak FP16 throughput.
Where does this NVIDIA A100 SXM4 pricing data come from?
Prices are read daily from each provider's public price list or API and recorded whenever they change. Listings are shown per GPU-hour on-demand unless marked otherwise, and the chart covers the last 30 days. The newest listing was checked 5h ago.
Why does the SXM4 cost less than the PCIe?
The SXM4 rents from $0.55 per GPU-hour, 9% less than the PCIe at $0.61. Only 1 listing tracks the PCIe, so its floor is one provider's choice rather than a market rate. The SXM4 is an SXM module, rated at 400 W against 250 W for the PCIe card. Hosting a module with that power and interconnect budget costs more than hosting a card, and the hourly rate carries it.