GPU MODELS / A10G

A10G cloud GPU pricing

7 listings across 4 providers. Cheapest published rate: $1.000/GPU/hr at Hugging Face. Sorted by per-GPU hourly price. Every row links to the provider's own published price.

WHAT IS THE A10G

The A10G is the AWS-specific variant of the A10, deployed in G5 instances with vGPU support. Same silicon, same 24 GB, same throughput - a mid-tier Ampere serving and workstation chip, priced by AWS's catalog rather than the spot GPU market.

MAKER
NVIDIA
ARCHITECTURE
Ampere
LAUNCHED
2021
MEMORY
24 GB GDDR6
MEMORY BANDWIDTH
600 GB/s
TENSOR COMPUTE
125 TFLOPS FP16 tensor dense (250 sparse)
POWER DRAW
150 W
INTERCONNECT
PCIe 4.0 x16

WHAT IT IS GOOD AT

Mid-size inference
7B-13B serving on AWS G5.
Graphics workstations
vGPU-licensed desktop and rendering workloads.

WHO SHOULD NOT RENT THE A10G

Non-AWS deployments - it is an AWS-specific SKU.
Workloads above 24 GB.
ProviderConfigVRAMLocation$/hr$/GPU/hrSource
Hugging FaceHugging FaceA10G × 124 GBUS$1.000$1.000receipt
AWSAWSA10G × 124 GBUS East (N. Virginia)$1.006$1.006receipt
BasetenBasetenA10G × 124 GBNot published$1.207$1.207receipt
Hugging FaceHugging FaceA10G × 424 GBUS$5.000$1.250receipt
AnyscaleAnyscaleA10G × 124 GBGlobal$1.363$1.363receipt
AWSAWSA10G × 424 GBUS East (N. Virginia)$5.672$1.418receipt
Hugging FaceHugging FaceA10G × 224 GBUS$3.000$1.500receipt

THE A10G MARKET, BY THE NUMBERS

PROVIDERS
4
CONFIGURATIONS
7
MEDIAN $/GPU/HR
$1.25
SPREAD
1.5x

4 providers publish 7 A10G configurations on the panel today. Published per-GPU hourly rates run from $1.000 (Hugging Face) to $1.50 (Hugging Face), a 2x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $1.25/GPU/hr. 6 rows were reported in stock by the provider at last observation; the rest are published catalog rates. All rates are on-demand - committed-use and reserved pricing is excluded.

A10G PRICE TREND, DAILY PANEL SNAPSHOTS

Across every A10G row on the panel, the daily floor moved from $1.207 to $1.000/GPU/hr over 10 days of tracking (-17.2%). Snapshots run daily; the full history is public in the repo.

A10G PROVIDER NOTES
Hugging Face · 3 configs · 1 region · 3 in stock
Sets the panel floor for this model.
from $1.000/hr
AWS · 2 configs · 1 region · 2 in stock
Within 15% of the panel floor.
from $1.006/hr
Baseten · 1 config
Published catalog rates.
from $1.207/hr
Anyscale · 1 config · 1 region · 1 in stock
1 of 1 configs reported in stock at last check.
from $1.363/hr
BUYING A10G: WHAT THE PANEL SAYS

At the current floor, one A10G running around the clock costs about $730 a month (730 hours of on-demand arithmetic). Per-GPU floors by config size: 1x $1.00 · 2x $1.50 · 4x $1.25. Compare the alternatives below before committing - the same budget often buys more than one chip class.

A10G VS THE ALTERNATIVES

A10G vs L4
AWS's own newer inference chip (G6 instances): similar compute, newer engines, often better $/hr. Check the panel's live AWS rows for both.

Compare live rates: A10G vs A10 · A10G vs L4

A10G PRICING FAQ

What is the cheapest A10G cloud GPU right now?
The cheapest published A10G rate on the panel is $1.000 per GPU-hour at Hugging Face (A10G x 1, US). Every price links to the provider's own published rate.
How much does a A10G cost per hour?
Across 4 providers tracked today, published per-GPU hourly rates for A10G run from $1.00 to $1.50, with a median of $1.25. Committed-use and reserved rates are excluded; everything shown is on-demand.
How many providers rent A10G GPUs?
4 providers publish 7 distinct A10G configurations on Compute Cafe. 6 of those rows were reported in stock by the provider's live endpoint at last observation.
Are A10G prices going up or down?
Over the 10 days Compute Cafe has tracked the panel, the A10G floor is falling: $1.207 to $1.000/GPU/hr (-17.2%). Daily snapshots are public in the repo, so the trend is auditable.
What does a A10G cost per month?
At the current floor of $1.000/GPU/hr, a A10G running 24/7 costs about $730 per month (730 hours of on-demand arithmetic on the cheapest published rate). Committed-use contracts can price lower; the panel tracks on-demand rates only.
Is A10G the same as A10?
Effectively yes - same GPU, tuned and licensed for AWS G5 instances. Any performance difference is instance configuration, not silicon.

See also alternatives to the A10G · cheapest GPU-hour per model · inference chips · training chips · by location