Blackwell server GPU for AI inference and visualization.
Weekly median price per GPU per hour · Get the data
By provider shows one card per company, placed where its best offer ranked. By configuration lists every offer.
What you can rent comes first: in stock, then waitlist, then not reported, then out of stock. Priced offers rank ahead of quote-only, and on-demand ahead of other billing types.
Within a group, five factors set the order:
We match the provider, its country, GPU model, form factor, billing type, availability and instance name. Matching is partial and case-insensitive. Everyday words work too, so "interruptible" finds spot and "sold out" finds out of stock.
No providers match .
Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Verify before provisioning. More on how we price.
One RTX PRO 6000 has 96 GB of VRAM. In practice, that's enough memory for roughly 146B parameters at 4-bit or 38B at 16-bit, assuming a 32K context. Below are some open-weight LLMs, with the estimated memory and GPUs each one needs.
| Model | Memory (INT4 / FP4) | RTX PRO 6000s needed | Cost /hr | Cost /mo |
|---|---|---|---|---|
Qwen3.8-27B
27B
|
18 GB
|
1
|
$2.19
|
$1,577
|
Gemma 4 31B
30.7B
|
22 GB
|
1
|
$2.19
|
$1,577
|
GPT-OSS-120B
117B · 5.1B active
|
67 GB
|
1
|
$2.19
|
$1,577
|
GLM-5.3-Flash
320B · 18B active
|
178 GB
|
4
|
$8.76
|
$6,307
|
MiniMax-M3
428B · 23B active
|
239 GB
|
4
|
$8.76
|
$6,307
|
DeepSeek V4.1 Flash
552B · 16B active
|
306 GB
|
4
|
$8.76
|
$6,307
|
Kimi K3
2.8T · 104B active
|
1,544 GB
|
18
|
–
|
–
|
Estimates based on the median on-demand rate. Memory is weights plus FP8 KV cache at 32K context per request (FP16/BF16 in the 16-bit column). GPU counts assume 90% of advertised VRAM is usable. No guarantee of runtime support, usable performance, or that a matching quantized build exists. Pricing methodology.
Nvidia RTX PRO 6000 · Per GPU · PCIe
| Compute · dense | |
|---|---|
| FP4 | 2,000 TFLOPS 4,000 with sparsity |
| FP8 | 1,000 TFLOPS 2,000 with sparsity |
| FP16 / BF16 | 500 TFLOPS 1,000 with sparsity |
| INT8 | 1,000 TOPS 2,000 with sparsity |
| FP32 | 120 TFLOPS |
| Precision support | FP4FP8FP16BF16TF32FP32INT8 |
| Memory | |
|---|---|
| Capacity | 96 GB GDDR7 |
| Bandwidth | 1,597 GB/s |
| Bus width | 512-bit |
| ECC | Yes |
| Silicon | |
|---|---|
| Architecture | Blackwell |
| Process | TSMC 4N |
| Transistors | 92.2 billion |
| Shader cores | 24,064 CUDA cores |
| Matrix cores | 752 Tensor cores |
| Compute units | 188 SMs |
| Fabric and host | |
|---|---|
| Host interface | PCIe 5.0 x16 |
| Power | |
|---|---|
| Board power | 600 W |
| Cooling | Passive |
As of September 28, 2026, the median on-demand price is $2.19 per GPU per hour across 41 providers with a priced on-demand config.
| Billing type | Configs | Median /GPU/hr | Cheapest |
|---|---|---|---|
On-demand | 193 | $2.19 | $0.65 |
Reserved | 180 | $1.89 | $0.48 (1 mo, in stock, HyperAI) |
Spot | 81 | $1.29 | $0.34 (in stock, Vast.ai) |
3 providers also quote the RTX PRO 6000 on a custom contract, priced per deal.
At 720 hours per month, one RTX PRO 6000 costs an estimated $1,577 at the median on-demand price. Cheapest verified in stock: $475 per month on-demand, $346 reserved (1 mo), $245 spot.
GetDeploying currently tracks RTX PRO 6000 configs from 53 providers. The cheapest verified in-stock on-demand configs come from Packet·ai, HyperAI, GPUhub and Lium. See the full price comparison above for every provider and config.
As of September 28, 2026, the median on-demand price has been flat over the past 90 days, at about $2.19 per GPU per hour, though it is about 22% above where it was a year ago.
One RTX PRO 6000 runs models up to roughly 146B parameters at 4-bit quantization or 38B at 16-bit, assuming a 32K context. Larger models run across multiple GPUs: the model table above shows the estimated memory and GPU count for popular open-weight LLMs.
96GB GDDR7 with Blackwell FP4 Tensor Cores. Fits very large models and datasets on a single GPU.
Nvidia H100
80 GB HBM3 · 2.1x the bandwidth of the RTX PRO 6000
Nvidia B200
180 GB HBM3e · 4.8x the bandwidth of the RTX PRO 6000
Nvidia H200
141 GB HBM3e · 3x the bandwidth of the RTX PRO 6000
Nvidia RTX 5090
32 GB GDDR7 · consumer card
Nvidia A100
80 GB HBM2e · 21% more bandwidth than the RTX PRO 6000
Last updated
