GPU MODELS / RTX A6000

RTX A6000 cloud GPU pricing

40 listings across 24 providers. Cheapest published rate: $0.320/GPU/hr at IO.NET. Sorted by per-GPU hourly price. Every row links to the provider's own published price.

WHAT IS THE RTX A6000

The RTX A6000 was the flagship workstation card of the Ampere generation and a budget-lab favorite: 48 GB of memory, full double-precision support, and widespread availability on budget GPU clouds. For fine-tuning and 34B-class inference it remains a solid mid-market option.

MAKER
NVIDIA
ARCHITECTURE
Ampere
LAUNCHED
2020
MEMORY
48 GB GDDR6 ECC
MEMORY BANDWIDTH
768 GB/s
TENSOR COMPUTE
155 TFLOPS FP16 tensor dense (310 sparse)
POWER DRAW
300 W
INTERCONNECT
PCIe 4.0 x16 (NVLink bridge on pairs)

WHAT IT IS GOOD AT

Fine-tuning 7B-13B
48 GB handles LoRA/QLoRA comfortably.
Inference to 34B
Quantized 34B fits per card.
Research workstations
FP32/FP64 support suits mixed workloads.

WHO SHOULD NOT RENT THE RTX A6000

High-throughput FP8 serving - no FP8 support.
Anything needing more than 48 GB per card.
ProviderConfigVRAMLocation$/hr$/GPU/hrSource
SpotGPUsSpotGPUsRTX A6000 × 148 GBNot published$0.170$0.170STALEreceipt
IO.NETIO.NETRTX A6000 × 148 GBNot published$0.320$0.320receipt
RunPodRunPodRTX A6000 × 148 GBGlobal$0.330$0.330receipt
Thunder ComputeThunder ComputeRTX A6000 × 148 GBNot published$0.350$0.350receipt
QuickPodQuickPodRTX A6000 × 148 GBNot published$0.350$0.350STALEreceipt
AI GalaxyAI GalaxyRTX A6000 × 148 GBChina$0.359$0.359receipt
MatPoolMatPoolRTX A6000 × 147.5 GBChina$0.426$0.426receipt
Yotta LabsYotta LabsRTX A6000 × 148 GBNot published$0.450$0.450receipt
TensorDockTensorDockRTX A6000 × 148 GBMultiple regions$0.470$0.470receipt
FeaturizeFeaturizeRTX A6000 × 151 GBChina$0.492$0.492receipt
HyperstackHyperstackA6000 × 148 GBUS / EU / CA$0.500$0.500receipt
HydraHostHydraHostRTX A6000 × 148 GBNot published$0.520$0.520receipt
gpu.aigpu.aiRTX A6000 × 148 GBNot published$0.540$0.540receipt
BeamBeamA6000 (from) × 148 GBNot published$0.540$0.540receipt
Massed ComputeMassed ComputeRTX A6000 [ALT Config] × 148 GBNot published$0.550$0.550receipt
OblivusOblivusRTX A6000 × 148 GBNot published$0.550$0.550receipt
Massed ComputeMassed ComputeRTX A6000 [ALT Config] × 848 GBNot published$4.400$0.550receipt
Massed ComputeMassed ComputeRTX A6000 × 148 GBNot published$0.570$0.570receipt
Massed ComputeMassed ComputeRTX A6000 [Premium] × 148 GBNot published$0.570$0.570receipt
Massed ComputeMassed ComputeRTX A6000 × 848 GBNot published$4.560$0.570receipt
Massed ComputeMassed ComputeRTX A6000 [Premium] × 848 GBNot published$4.560$0.570receipt
Hostnot GPUHostnot GPURTX A6000 × 248 GBUS Central$1.190$0.595receipt
GPUBrazilGPUBrazilRTX A6000 × 148 GBAsia-Pacific$0.596$0.596receipt
GPUBrazilGPUBrazilRTX A6000 × 148 GBEurope$0.596$0.596receipt
GPUBrazilGPUBrazilRTX A6000 × 148 GBNorth America$0.596$0.596receipt
Hostnot GPUHostnot GPURTX A6000 × 148 GBUS Central$0.600$0.600receipt
Hostnot GPUHostnot GPURTX A6000 × 448 GBEU East$2.550$0.637receipt
Hostnot GPUHostnot GPURTX A6000 × 248 GBUS East$1.410$0.705receipt
Hostnot GPUHostnot GPURTX A6000 × 148 GBUS East$0.710$0.710receipt
GPUBrazilGPUBrazilRTX A6000 × 148 GBGlobal$0.782$0.782receipt
Hostnot GPUHostnot GPURTX A6000 × 148 GBEU West$0.790$0.790receipt
SeewebSeewebRTX A6000 × 148 GBItaly$0.822$0.822STALEreceipt
Verda (formerly DataCrunch)Verda (formerly DataCrunch)RTX A6000 × 148 GBEurope (Finland / Iceland)$0.861$0.861receipt
Verda (formerly DataCrunch)Verda (formerly DataCrunch)RTX A6000 × 248 GBEurope (Finland / Iceland)$1.722$0.861receipt
Verda (formerly DataCrunch)Verda (formerly DataCrunch)RTX A6000 × 448 GBEurope (Finland / Iceland)$3.444$0.861receipt
Verda (formerly DataCrunch)Verda (formerly DataCrunch)RTX A6000 × 848 GBEurope (Finland / Iceland)$6.888$0.861receipt
ACCACCRTX A6000 × 148 GBSingapore$1.220$1.220receipt
InHostedInHostedRTX A6000 × 148 GBIndia$1.256$1.256receipt
QarnotQarnotRTX A6000 × 248 GBEU$2.989$1.495STALEreceipt
PaperspacePaperspaceA6000 × 148 GBNot published$1.890$1.890receipt

THE RTX A6000 MARKET, BY THE NUMBERS

PROVIDERS
24
CONFIGURATIONS
40
MEDIAN $/GPU/HR
$0.57
SPREAD
5.9x

24 providers publish 40 RTX A6000 configurations on the panel today. Published per-GPU hourly rates run from $0.320 (IO.NET) to $1.89 (Paperspace), a 6x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $0.57/GPU/hr. 11 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.

CHEAPEST RTX A6000 BY GEOGRAPHY
chinaAI Galaxy$0.359/hrunited statesHyperstack$0.500/hrcanadaHyperstack$0.500/hrfinlandVerda (formerly DataCrunch)$0.861/hricelandVerda (formerly DataCrunch)$0.861/hrsingaporeACC$1.220/hrindiaInHosted$1.256/hr
RTX A6000 PRICE TREND, DAILY PANEL SNAPSHOTS

Across every RTX A6000 row on the panel, the daily floor moved from $0.330 to $0.320/GPU/hr over 53 days of tracking (-3.0%). The median continuously-tracked listing held flat at $0.57. Snapshots run daily; the full history is public in the repo.

RTX A6000 PROVIDER NOTES
IO.NET · 1 config
Sets the panel floor for this model.
from $0.320/hr
RunPod · 1 config · 1 region
Within 15% of the panel floor.
from $0.330/hr
Thunder Compute · 1 config · 1 in stock
Within 15% of the panel floor.
from $0.350/hr
AI Galaxy · 1 config · 1 region
Within 15% of the panel floor.
from $0.359/hr
MatPool · 1 config · 1 region
Published catalog rates.
from $0.426/hr
Yotta Labs · 1 config
Published catalog rates.
from $0.450/hr
TensorDock · 1 config · 1 region
Published catalog rates.
from $0.470/hr
Featurize · 1 config · 1 region
Published catalog rates.
from $0.492/hr
BUYING RTX A6000: WHAT THE PANEL SAYS

At the current floor, one RTX A6000 running around the clock costs about $234 a month (730 hours of on-demand arithmetic). Node shape matters: the single-GPU floor is $0.320/GPU/hr against $0.550/GPU/hr on an 8-GPU node (+71.9%). Per-GPU floors by config size: 1x $0.32 · 2x $0.59 · 4x $0.64 · 8x $0.55. Compare the alternatives below before committing - the same budget often buys more than one chip class.

RTX A6000 VS THE ALTERNATIVES

RTX A6000 vs RTX 6000 Ada
The Ada successor adds FP8, ~2.3x tensor throughput and 25% more bandwidth at the same 48 GB and 300 W. Newer is better unless the A6000 is much cheaper.
RTX A6000 vs L40S
L40S is faster at FP16 with FP8 support; A6000 counters with FP64 and sometimes lower panel prices. For pure AI serving, L40S; for mixed precision work, compare rates.

Compare live rates: RTX A6000 vs RTX 6000 Ada · RTX A6000 vs L40S

RTX A6000 PRICING FAQ

What is the cheapest RTX A6000 cloud GPU right now?
The cheapest published RTX A6000 rate on the panel is $0.320 per GPU-hour at IO.NET (RTX A6000 x 1, Not published). Every price links to the provider's own published rate.
How much does a RTX A6000 cost per hour?
Across 24 providers tracked today, published per-GPU hourly rates for RTX A6000 run from $0.32 to $1.89, with a median of $0.57. Committed-use and reserved rates are excluded; everything shown is on-demand.
How many providers rent RTX A6000 GPUs?
24 providers publish 40 distinct RTX A6000 configurations on Compute Cafe. 11 of those rows were reported in stock by the provider's live endpoint at last observation.
Where is RTX A6000 cheapest?
By published per-GPU hourly rate, the cheapest geography for RTX A6000 today is china at $0.36/hr (AI Galaxy). The spread across 7 tracked geographies runs to $1.26/hr in india.
Are RTX A6000 prices going up or down?
Over the 53 days Compute Cafe has tracked the panel, the RTX A6000 floor is falling: $0.330 to $0.320/GPU/hr (-3.0%). The median continuously-tracked listing moved from $0.57 to $0.57 (+0.0%). Daily snapshots are public in the repo, so the trend is auditable.
What does a RTX A6000 cost per month?
At the current floor of $0.320/GPU/hr, a RTX A6000 running 24/7 costs about $234 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 a full 8-GPU RTX A6000 node cheaper than a single card?
On today's panel: the single-GPU floor is $0.320/GPU/hr and the 8-GPU node floor is $0.550/GPU/hr - more expensive per GPU on a full node (+71.9%). Check the table for the exact configs behind each floor.
Is the RTX A6000 still good for ML?
Yes for fine-tuning and mid-size inference - 48 GB is the draw. It lacks FP8, so for high-throughput serving an L40S or RTX 6000 Ada is faster.

See also alternatives to the RTX A6000 · cheapest GPU-hour per model · inference chips · training chips · by location