Is the RTX PRO 6000 WS suitable for large language model inference?

Yes. With 96 GB of VRAM and about 1403 GB/s of memory bandwidth, this workstation GPU can hold substantial model weights and context windows in memory. You can deploy inference servers quickly using the provided vLLM or Ollama templates.

How does this GPU compare to consumer or datacenter options?

As a Blackwell workstation GPU, the RTX PRO 6000 WS offers a 96 GB VRAM capacity that exceeds standard consumer cards, allowing larger workloads to fit on a single card. Datacenter GPUs may offer higher memory bandwidth, whereas smaller workstation or consumer GPUs are sufficient for lighter tasks.

What software environments are available for this GPU?

Instances support compute capability 12.0 and come with ready-to-use templates including PyTorch/CUDA, vLLM, Ollama, and ComfyUI. Full SSH access is included so you can customize your software stack as needed.

How does billing work on Cloudj?

Rentals on Cloudj use per-hour billing, so you only pay for the time your instance is active. You can connect via SSH to run your tasks and terminate the instance when finished.

GPU Cloud · NVIDIA · Blackwell

Rent NVIDIA RTX PRO 6000 WS GPU by the hour

The NVIDIA RTX PRO 6000 WS is a Blackwell-architecture workstation GPU available for on-demand cloud rental on Cloudj. Equipped with 96 GB of VRAM and a 600 W power envelope, it delivers about 119.0 TFLOPS of standard floating-point performance and about 1403 GB/s of memory bandwidth. It supports compute capability 12.0 for modern AI and compute workloads.

Rent RTX PRO 6000 WS All GPUs

Specifications

VRAM per GPU

96 GB

Architecture

NVIDIA Blackwell

Compute capability

FP32 compute

≈ 119 TFLOPS

Memory bandwidth

≈ 1403 GB/s

Power limit

600 W

Values reported by the infrastructure provider for machines currently listed; real performance varies by machine.

Available configurations

Live stock and prices, updated continuously. The final price is confirmed before you order.

Rent now

What the RTX PRO 6000 WS is good for

Large Language Model Inference

Run large language models in a single 96 GB memory pool using pre-configured vLLM or Ollama templates.

Model Fine-Tuning and Training

Train and fine-tune deep learning models with PyTorch/CUDA templates, leveraging about 119.0 TFLOPS of compute.

High-Resolution Generative Media

Generate complex image and video outputs in ComfyUI without hitting memory bottlenecks using 96 GB of VRAM.

Workstation Compute and Rendering

Execute heavy workstation simulations and rendering tasks backed by about 1403 GB/s of memory bandwidth.

Is it the right GPU for you?

Choose the RTX PRO 6000 WS when your workloads require 96 GB of VRAM and Blackwell architecture features without moving to datacenter-segment hardware. If your models require less memory, a consumer or smaller workstation GPU is more economical, while extreme-scale training jobs may require a datacenter-class GPU with higher memory bandwidth.

Compare with similar GPUs

GPU	VRAM per GPU	FP32 compute	Power limit
NVIDIA RTX PRO 6000 WS	96 GB	≈ 119 TFLOPS	600 W
NVIDIA RTX PRO 6000 Max-Q	96 GB	≈ 119 TFLOPS	300 W	Details →
NVIDIA RTX PRO 6000 S	96 GB	≈ 93.6 TFLOPS	600 W	Details →
NVIDIA RTX 6000D	84 GB	≈ 77.6 TFLOPS	600 W	Details →
NVIDIA RTX 6000Ada	48 GB	≈ 81.4 TFLOPS	300 W	Details →

Frequently asked questions

Rent the RTX PRO 6000 WS by the hour

Create an account, choose a configuration and connect over SSH. Billing stops when you delete the GPU.

Rent RTX PRO 6000 WS All GPUs