Rent RTX 4090 GPU Cloud
NVIDIA RTX 4090 24GB on-demand from $1.00/hr. Per-second billing, scale to zero, no idle costs. Deploy your first workload in under 60 seconds.
Best for RTX 4090 workloads
LLM inference
High CUDA core count makes the 4090 excellent for fast token generation. Run 7–13B models with low latency at $1.00/hr.
Image generation
1 TB/s+ memory bandwidth enables fast Stable Diffusion and SDXL generation. Ideal for high-throughput image workloads.
Fine-tuning with LoRA/QLoRA
24 GB VRAM is enough for fine-tuning 7–13B models with quantization. Pay per second for training runs — no idle cost.
RTX 4090 specifications
| GPU | NVIDIA RTX 4090 24GB |
| VRAM | 24 GB |
| Architecture | Ada Lovelace |
| CUDA Cores | 16,384 |
| Tensor Cores | 512 |
| Memory Bandwidth | 1,008 GB/s |
RTX 4090 pricing on Velar
Serverless Jobs
Scale to zero · per-second billing
$1.00/hr
- $0.000278/sec — billed to the second
- Scale to zero when idle
- No reserved capacity
Persistent Endpoint
Always-on · flat monthly rate
$600/mo
- $0.83/hr effective — 17% cheaper than on-demand
- Zero cold-start latency
- Always available, guaranteed GPU
Pro plan required · See plans
Deploy on RTX 4090 in 12 lines of Python
No Dockerfile. No Kubernetes. Just a Python decorator and a GPU string.
import velar
app = velar.App("rtx-inference")
image = velar.Image.from_registry(
"pytorch/pytorch:2.1.0-cuda12.1-cudnn8-runtime"
).pip_install("transformers", "accelerate")
@app.function(gpu="RTX4090", image=image)
def generate(prompt: str) -> str:
from transformers import pipeline
pipe = pipeline("text-generation", model="mistralai/Mistral-7B-v0.1")
return pipe(prompt, max_new_tokens=256)[0]["generated_text"]
app.deploy()
print(generate.remote("Explain transformers in one paragraph"))Also available on Velar
Deploy on RTX 4090 today
Start with $10 in free GPU credits. No credit card required. First workload live in under 60 seconds.