Use cases
Recommended GPU classes per workload and estimated hourly cost ranges derived from live catalog prices.
Step-by-step guides
Run Stable Diffusion on an RTX 4090 in the cloudRTX 4090 · RTX 5090Run ComfyUI on a cloud GPURTX 4090 · RTX 5090Run Llama and other open LLMs on a cloud GPURTX 4090 · RTX 5090 · A100 SXM4Fine-tune an LLM with LoRA on a cloud GPUA100 SXM4 · RTX 5090Jupyter notebooks on a cloud GPURTX 4090 · RTX 5090 · A100 SXM4Transcribe audio with Whisper on a cloud GPURTX 4090 · RTX 3090Generate AI videos on a cloud GPURTX 5090 · A100 SXM4 · H100 SXMGenerate AI music on a cloud GPURTX 4090 · RTX 3090Generate AI images with Flux on a cloud GPURTX 4090 · RTX 5090
LLM training & fine-tuning
High-VRAM datacenter cards for training large language models and LoRA/QLoRA fine-tuning. H100 class with NVLink for multi-node training; A100 is enough on a single node.
Recommended GPU
A100 / H100
Estimated hourly cost
$1.50–4.50/hr
Ranges vary with current catalog prices; check the catalog for exact pricing.
Go to catalog