GPU Cloud Pricing · Blackwell
The fastest consumer GPU ever shipped: 32 GB of GDDR7 at 1.8 TB/s, and the best raw price/performance for single-GPU inference today.
GPU.ai rates are the live fleet floor — the cheapest provider with stock, passed through at cost. Competitor rates are public on-demand list prices.
| Provider | $ / GPU / hr |
|---|---|
| GPU.ai — Secure Cloud | $0.990 |
| GPU.ai — Community Cloud | $0.295 |
| CoreWeave | Not available |
| AWS | Not available |
| Azure | Not available |
Need scale? Real 4× RTX 5090 nodes are in inventory from $0.295/GPU/hr.
Available regions: Asia Pacific · Canada Central · Europe East · Europe West · US East
GPU.ai prices are read live from our aggregated fleet of 12+ GPU clouds and normalized to a per-GPU hourly rate (a 4× node's price is divided by its GPU count). We charge the provider's raw rate with zero markup and bill per second. Competitor figures are each provider's public on-demand list rate for the closest equivalent SKU — never spot, reserved, or committed-use discounts. Stock counts reflect launchable inventory, not theoretical capacity. Last updated 2026-08-17; this page refreshes hourly.
| Architecture | Blackwell |
| VRAM | 32 GB GDDR7 |
| Memory bandwidth | 1,792 GB/s |
| FP16 tensor (dense) | — |
| CUDA cores | 21,760 |
| Power (TDP) | 575 W |
| Interconnect | PCIe 5.0 |
| Released | 2025 |
Minimum RTX 5090 count needed to hold popular open models in VRAM, at FP16 and 4-bit quantization.
| Model | Params | FP16 | 4-bit |
|---|---|---|---|
| Llama 3.1 8B | 8B | 1× RTX 5090 | 1× RTX 5090 |
| Mistral 7B | 7B | 1× RTX 5090 | 1× RTX 5090 |
| Gemma 2 27B | 27B | 4× RTX 5090 | 1× RTX 5090 |
| Qwen2.5 32B | 32B | 4× RTX 5090 | 1× RTX 5090 |
| Llama 3.1 70B | 70B | 8× RTX 5090 | 2× RTX 5090 |
| Qwen2.5 72B | 72B | 8× RTX 5090 | 2× RTX 5090 |
| Mixtral 8x22B | 141B MoE | Doesn't fit | 4× RTX 5090 |
| Llama 3.1 405B | 405B | Doesn't fit | Doesn't fit |
| FLUX.1 dev | 12B | 1× RTX 5090 | 1× RTX 5090 |
| SDXL | 3.5B | 1× RTX 5090 | 1× RTX 5090 |
| Whisper large-v3 | 1.5B | 1× RTX 5090 | 1× RTX 5090 |
Weights-only estimates with a 20% margin for KV cache and activations at moderate context. Long contexts, large batches, and training need substantially more memory.
RTX 5090 cloud instances on GPU.ai currently start at $0.295 per GPU per hour, billed per second. That is the live floor price across every provider in our fleet — we route your launch to the cheapest provider with stock and add zero markup.
The NVIDIA GeForce RTX 5090 has 32 GB of GDDR7 memory with up to 1,792 GB/s of bandwidth. GPU.ai reports the minimum VRAM actually delivered across our fleet, so the figure you see at launch is the figure you get.
Yes — real 4× RTX 5090 nodes are launchable right now, from $0.295 per GPU per hour. We only advertise node sizes that exist in live inventory, not theoretical configurations.
GPU.ai aggregates a dozen-plus GPU clouds and passes through each provider's raw price with no markup, so the same silicon is typically far cheaper than hyperscaler list rates. The comparison table on this page shows the current public on-demand list prices side by side.
Billing is per second, starting when your instance is SSH-ready and stopping the moment you terminate. There are no minimum commitments, no reservation fees, and no egress surprises — you pay the listed hourly rate pro-rated to the second.