Rent an NVIDIA RTX PRO 6000 Blackwell from $0.19 per GPU-hour
NVIDIA RTX PRO 6000 Blackwell cloud pricing: $0.19 per GPU-hour on spot, $0.39 on-demand, $0.29 reserved, billed by the minute. Blackwell in a workstation board with 96 GB of GDDR7 — more memory than an H100 SXM, at workstation prices. Excellent for single-GPU inference of large models and for rendering.
Billed per started minute. 768 GB of GPU memory on a 8× node. Persistent storage $0.08 per GB-month. No egress fees.
- GPU memory
- 96GBGDDR7 · MIG capable
- Bandwidth
- 1,597GB/sSets token throughput on inference
- Node sizes
- 1× · 2× · 4× · 8×Same price per GPU at every size
- Reclaim rate
- 5–10%Spot instances reclaimed over 30 days
What you get with each RTX PRO 6000 node.
| Node | GPU memory | vCPU | RAM | Local NVMe | Spot | On-demand | Action |
|---|---|---|---|---|---|---|---|
| 1× RTX PRO 6000 | 96 GB | 16 | 128 GB | 1 TB | $0.19/h | $0.39/h | Configure |
| 2× RTX PRO 6000 | 192 GB | 32 | 256 GB | 2 TB | $0.38/h | $0.78/h | Configure |
| 4× RTX PRO 6000 | 384 GB | 64 | 512 GB | 4 TB | $0.76/h | $1.56/h | Configure |
| 8× RTX PRO 6000 | 768 GB | 128 | 1,024 GB | 8 TB | $1.52/h | $3.12/h | Configure |
GPU-to-GPU link: PCIe Gen5 x16. Local NVMe is scratch space wiped when the instance ends; keep anything you need on a persistent disk.
48 RTX PRO 6000 GPUs free right now.
Straight from the capacity pool the console books against. Prices are identical in every region — pick the one closest to your data.
A region with zero free GPUs still accepts reserved capacity requests — we hold hardware for a term rather than sell what is already taken.
What fits in 96 GB — and what it costs per hour.
Weights need about 2 GB per billion parameters at FP16 / BF16, 1 GB per billion parameters at FP8, 0.5 GB per billion parameters at INT4. We keep 20% of the memory free for activations, the KV cache and the CUDA context, then take the smallest node that still fits.
| Open model | Parameters | FP16 / BF16node · spot price | FP8node · spot price | INT4node · spot price |
|---|---|---|---|---|
| Llama 3.1 8BThe default small model for fine-tuning and agents | 8B | 1× $0.19/h | 1× $0.19/h | 1× $0.19/h |
| Qwen3 14BMid-size dense Qwen3 | 14.8B | 1× $0.19/h | 1× $0.19/h | 1× $0.19/h |
| gpt-oss-20bSmall open-weight reasoning model for local and edge serving | 21B | 1× $0.19/h | 1× $0.19/h | 1× $0.19/h |
| Gemma 4 31BLargest dense Gemma 4, multimodal, 256K context | 31.3B | 1× $0.19/h | 1× $0.19/h | 1× $0.19/h |
| Qwen3.6 35B A3BSmall hybrid-attention mixture of experts | 35B | 1× $0.19/h | 1× $0.19/h | 1× $0.19/h |
| Llama 3.3 70BThe common 70B baseline (same size and layout as Llama 3.1 70B) | 70.6B | 2× $0.38/h | 1× $0.19/h | 1× $0.19/h |
| gpt-oss-120bOpen-weight reasoning model, mixture of experts, tool use | 117B | 4× $0.76/h | 2× $0.38/h | 1× $0.19/h |
| Qwen3 235B A22BFlagship Qwen3 mixture of experts, the 2507 checkpoint | 235B | 8× $1.52/h | 4× $0.76/h | 2× $0.38/h |
| DeepSeek R1Reasoning model, the R1-0528 checkpoint | 671B | over 8× | over 8× | 8× $1.52/h |
Each model links to its full sizing: KV cache by context length and every GPU that runs it — GPU requirements for 64 open models.
Real usage depends on context length, batch size and the serving engine: a long context can add tens of gigabytes of KV cache. Treat the table as the floor, not the ceiling — and if a job is close to the limit, take the next node size or quantise one step further.
What a dollar buys on an RTX PRO 6000.
Ranked 4 of 12 workstation models in this catalogue on cost per gigabyte of GPU memory.
GPU memory you get for $1.00 of spot time, per hour.
Bandwidth decides token throughput far more often than raw FLOPS.
Spot price divided by published FP32 throughput.
One GPU on spot, billed per minute, storage excluded.
Against the rest of the market
68 providers · September 2026- Median on-demand elsewhere
- $1.90Our on-demand is 79% below it
- Cheapest on-demand seen
- $0.54Our on-demand $0.39 is 28% below it
- Cheapest spot seen
- $0.30Our spot $0.19 is 37% below it
- 24 hours on one GPU
- $4.56 on spot$9.36 on-demand · $12.96 at the cheapest rate found elsewhere
Index built from published prices for the same GPU across 102 providers (provider pricing pages and public price indexes), September 2026. How the guarantee is enforced.
RTX PRO 6000 Blackwell price per hour at 58 other providers.
The lowest published on-demand rate we recorded at each provider, cheapest first: $0.59 to $7.94 per GPU-hour, with a median of $1.92. Every row links to its source; a provider without a published price is left out rather than estimated.
| # | Provider | On-demandper GPU-hour | Spot | 730 h | Source |
|---|---|---|---|---|---|
| — | OnlineGPUthis site · how the price is set | $0.39 | $0.19 | $284.70 | pricing page |
| 1 | GPU Outletrecorded 15 Sep 2026 | $0.59 | not published | $430.70 | computeprices.com |
| 2 | RunPodrecorded 15 Sep 2026 | $0.59 | not published | $430.70 | runpod.io |
| 3 | Packet.airecorded 15 Sep 2026 | $0.66 | not published | $481.80 | getdeploying.com |
| 4 | HyperAIrecorded 15 Sep 2026 | $0.80 | not published | $584.00 | getdeploying.com |
| 5 | Liumrecorded 15 Sep 2026 | $0.82 | not published | $598.60 | getdeploying.com |
| 6 | Vast.airecorded 15 Sep 2026 | $0.87 | $0.34 | $635.10 | getdeploying.com |
| 7 | Database Martrecorded 15 Sep 2026 | $0.90 | not published | $657.00 | getdeploying.com |
| 8 | Leasewebrecorded 15 Sep 2026 | $0.90 | not published | $657.00 | getdeploying.com |
| 9 | Verdarecorded 15 Sep 2026 | $0.92 | $0.92 | $672.33 | shadeform.com |
| 10 | Nebiusrecorded 15 Sep 2026 | $0.95 | $0.95 | $693.50 | nebius.com |
| 11 | GPUhubrecorded 15 Sep 2026 | $0.97 | not published | $708.10 | getdeploying.com |
| 12 | Boost Runrecorded 15 Sep 2026 | $0.99 | not published | $722.70 | shadeform.com |
| 13 | Beamrecorded 15 Sep 2026 | $1.09 | not published | $795.70 | getdeploying.com |
| 14 | Nova Cloudrecorded 15 Sep 2026 | $1.14 | $0.30 | $832.20 | cloud-gpus.com |
| 15 | Spheronrecorded 15 Sep 2026 | $1.17 | $1.16 | $854.10 | getdeploying.com |
Show the other 43 providers
| # | Provider | On-demandper GPU-hour | Spot | 730 h | Source |
|---|---|---|---|---|---|
| 16 | 1Legionrecorded 15 Sep 2026 | $1.19 | not published | $868.70 | computeprices.com |
| 17 | CoreWeaverecorded 15 Sep 2026 | $1.20 | $1.20 | $876.00 | coreweave.com |
| 18 | Hyperstackrecorded 15 Sep 2026 | $1.30 | $1.48 | $949.00 | hyperstack.cloud |
| 19 | Cirrascalerecorded 15 Sep 2026 | $1.39 | not published | $1,014.70 | getdeploying.com |
| 20 | Seewebrecorded 15 Sep 2026 | $1.44 | not published | $1,051.20 | computeprices.com |
| 21 | GPU.airecorded 15 Sep 2026 | $1.47 | not published | $1,073.10 | getdeploying.com |
| 22 | Uthorecorded 15 Sep 2026 | $1.69 | not published | $1,233.70 | getdeploying.com |
| 23 | Google Cloudrecorded 15 Sep 2026 | $1.72 | $1.74 | $1,255.60 | cloud.google.com |
| 24 | Akash Networkrecorded 15 Sep 2026 | $1.76 | not published | $1,284.80 | console.akash.network |
| 25 | EmpirioLabs AIrecorded 15 Sep 2026 | $1.80 | not published | $1,314.00 | getdeploying.com |
| 26 | SwissGPUrecorded 15 Sep 2026 | $1.84 | not published | $1,343.20 | getdeploying.com |
| 27 | EcoHashrecorded 15 Sep 2026 | $1.89 | not published | $1,379.70 | getdeploying.com |
| 28 | Jarvis Labsrecorded 15 Sep 2026 | $1.89 | $0.99 | $1,379.70 | computeprices.com |
| 29 | Latitude.shrecorded 15 Sep 2026 | $1.90 | not published | $1,387.00 | latitude.sh |
| 30 | Cloudzyrecorded 15 Sep 2026 | $1.93 | not published | $1,408.90 | getdeploying.com |
| 31 | Envergerecorded 15 Sep 2026 | $1.95 | not published | $1,423.50 | getdeploying.com |
| 32 | IO.NETrecorded 15 Sep 2026 | $1.99 | not published | $1,452.70 | computeprices.com |
| 33 | AceCloudrecorded 15 Sep 2026 | $2.00 | not published | $1,460.00 | getdeploying.com |
| 34 | Exoscalerecorded 15 Sep 2026 | $2.00 | not published | $1,460.00 | getdeploying.com |
| 35 | Geoddrecorded 15 Sep 2026 | $2.14 | not published | $1,562.20 | getdeploying.com |
| 36 | Massed Computerecorded 15 Sep 2026 | $2.19 | not published | $1,597.97 | vm.massedcompute.com |
| 37 | Shadeformrecorded 15 Sep 2026 | $2.19 | not published | $1,598.70 | computeprices.com |
| 38 | Koyebrecorded 15 Sep 2026 | $2.20 | not published | $1,606.00 | getdeploying.com |
| 39 | Hetznerrecorded 15 Sep 2026 | $2.22 | not published | $1,620.60 | getdeploying.com |
| 40 | Contaborecorded 15 Sep 2026 | $2.25 | not published | $1,642.50 | getdeploying.com |
| 41 | Omega Gradientrecorded 15 Sep 2026 | $2.30 | not published | $1,679.00 | computeprices.com |
| 42 | Runcraterecorded 15 Sep 2026 | $2.41 | not published | $1,759.30 | getdeploying.com |
| 43 | Sestercerecorded 15 Sep 2026 | $2.41 | not published | $1,759.30 | getdeploying.com |
| 44 | Cerebriumrecorded 15 Sep 2026 | $2.50 | not published | $1,825.00 | getdeploying.com |
| 45 | AtmosComputerecorded 15 Sep 2026 | $2.51 | not published | $1,832.30 | computeprices.com |
| 46 | Oblivusrecorded 15 Sep 2026 | $2.60 | not published | $1,898.00 | getdeploying.com |
| 47 | Cyfuture AIrecorded 15 Sep 2026 | $2.68 | not published | $1,956.40 | getdeploying.com |
| 48 | Hugging Facerecorded 15 Sep 2026 | $2.75 | not published | $2,007.50 | getdeploying.com |
| 49 | Elastxrecorded 15 Sep 2026 | $2.89 | not published | $2,109.70 | getdeploying.com |
| 50 | Fal.airecorded 15 Sep 2026 | $2.99 | not published | $2,182.70 | getdeploying.com |
| 51 | Akamai Cloudrecorded 15 Sep 2026 | $3.00 | not published | $2,190.00 | getdeploying.com |
| 52 | Modalrecorded 15 Sep 2026 | $3.03 | not published | $2,211.90 | computeprices.com |
| 53 | Prime Intellectrecorded 15 Sep 2026 | $3.14 | not published | $2,292.20 | primeintellect.ai |
| 54 | Leafcloudrecorded 15 Sep 2026 | $3.19 | not published | $2,328.70 | getdeploying.com |
| 55 | AWSrecorded 15 Sep 2026 | $3.36 | not published | $2,452.80 | getdeploying.com |
| 56 | Oracle Cloudrecorded 15 Sep 2026 | $4.50 | not published | $3,285.00 | getdeploying.com |
| 57 | Microsoft Azurerecorded 15 Sep 2026 | $5.50 | $1.02 | $4,015.00 | getdeploying.com |
| 58 | Hinoderecorded 15 Sep 2026 | $7.94 | not published | $5,796.20 | computeprices.com |
Recorded 15 September 2026 from published price pages (provider pricing pages and public price indexes), compute only: storage, egress and taxes excluded. Prices move — if a row is out of date, send us the link and it is corrected.
What people run on an RTX PRO 6000 Blackwell.
Training and full fine-tuning
Enough memory for optimiser states and activations on models a smaller card can only run in inference.
Serving large models
A 70B model at FP8 across two GPUs, with room for a long context window.
Partitioned serving (MIG)
The GPU can be split into isolated instances, each with its own memory and compute slice.
Anything that can checkpoint
On the spot tier this model is 51% below its own on-demand price, with a 2-minute notice before a reclaim.
NVIDIA RTX PRO 6000 Blackwell, as published by NVIDIA.
Taken from the vendor datasheet. Where a figure is not published, the row is absent rather than estimated.
- Vendor
- NVIDIA
- Architecture
- Blackwell (2025)
- Segment
- Workstation
- Memory
- 96 GB GDDR7
- Memory bandwidth
- 1,597 GB/s
- FP32
- 120 TFLOPS
- CUDA cores
- 24,064
- Board power
- 600 W
- Form factor
- PCIe dual-slot
- Interconnect
- PCIe Gen5 x16
- Multi-instance (MIG)
- Supported by the GPU
- 8-bit float (FP8)
- Supported by the architecture
Source: www.nvidia.com · not confirmed on a vendor document
Software that runs on it
8 templatesPre-built environments, pulled on the node before you land on it. Or bring any OCI image from a public or private registry.
On the spot tier
5–10% reclaimed · 30 d- 2 minutes of notice on the metadata endpoint, a webhook and the console
- The instance is stopped, not deleted — disk and IP stay attached
- Auto-relaunch on the next free RTX PRO 6000, or switch the same disk to on-demand
- You save $4.80 a day per GPU against on-demand
If the RTX PRO 6000 is not the right fit.
Next step up in memory: 141 GB instead of 96 GB
RTX PRO 5000 Blackwell 48 GB GDDR7 · Blackwell $0.15/h spotSame Blackwell generation
RTX PRO 4500 Blackwell 32 GB GDDR7 · Blackwell $0.09/h spotSame Blackwell generation
RTX PRO 4000 Blackwell 24 GB GDDR7 · Blackwell $0.08/h spotSame Blackwell generation
Or put two of them side by side: RTX PRO 6000 vs H200 NVL · RTX PRO 6000 vs RTX PRO 5000 · RTX PRO 6000 vs RTX PRO 4500
RTX PRO 6000 Blackwell — the questions that come up
How much does an RTX PRO 6000 Blackwell cost per hour?
Spot is $0.19 per GPU-hour and on-demand is $0.39, both billed per minute — every started minute costs the hourly price divided by 60, so $0.0032 on spot. Reserved capacity is $0.29 per GPU-hour on a monthly commitment. Storage and egress are not included in that rate: persistent disks are $0.08 per GB-month and there are no egress fees.
How many GPUs can I put in one instance?
Node sizes are 1×, 2×, 4×, 8× — up to 8 RTX PRO 6000 Blackwell GPUs in a single virtual machine, with CPU, RAM and local NVMe scaled with the GPU count. The price per GPU is identical at every node size.
Can it run a 70B model?
Yes. At FP8, a 70B model needs about 70 GB for the weights, so it fits on 1× RTX PRO 6000 Blackwell ($0.19 per hour on spot) with roughly 20% of the memory left for activations and the KV cache.
Which regions have it?
Available now in US West (Oregon), EU Central (Frankfurt). Prices are identical in every region; pick the one closest to your data.
How often is a spot RTX PRO 6000 reclaimed?
Over the trailing 30 days, 5–10% of spot instances on this model were reclaimed. You get a 2-minute notice on the metadata endpoint, through a webhook and in the console; the instance is then stopped, never deleted, and the persistent disk stays attached.
Anything else about this model? Ask an engineer — the same people run the nodes. See also all prices, how we compare with 41 other providers and how to pay in crypto.
An RTX PRO 6000 Blackwell for $0.19 an hour, running in under 60 seconds.
Prepaid in crypto, pay as you go: no card, no contract, no minimum commitment. Add credit, launch, stop whenever you want.
Billed per minute from the moment the instance is reachable. Minimum credit $30, no subscription. 48 RTX PRO 6000 GPUs available right now.