Rent L4 24GB GPU Cloud
NVIDIA L4 24GB on-demand from $0.66/hr. Per-second billing, scale to zero, no idle costs. Deploy your first workload in under 60 seconds.
Best for L4 24GB workloads
Image & video generation
Run Stable Diffusion, SDXL, and video models. The L4 delivers great throughput for image workloads at the lowest cost per image on Velar.
Small LLM inference
Deploy Llama 3 8B, Mistral 7B, Qwen 2.5 7B, and similar models with vLLM or TGI. Enough VRAM for 7–13B parameter models in fp16.
Batch embeddings & transcription
Process large datasets cost-efficiently. Run Whisper, text embedding models, or custom inference pipelines at $0.000183/sec.
L4 24GB specifications
| GPU | NVIDIA L4 24GB |
| VRAM | 24 GB |
| Architecture | Ada Lovelace |
| CUDA Cores | 7,680 |
| Tensor Cores | 240 |
| Memory Bandwidth | 300 GB/s |
L4 24GB pricing on Velar
Serverless Jobs
Scale to zero · per-second billing
$0.66/hr
- $0.000183/sec — billed to the second
- Scale to zero when idle
- No reserved capacity
Deploy on L4 24GB in 12 lines of Python
No Dockerfile. No Kubernetes. Just a Python decorator and a GPU string.
import velar
app = velar.App("l4-inference")
image = velar.Image.from_registry(
"pytorch/pytorch:2.1.0-cuda12.1-cudnn8-runtime"
).pip_install("transformers", "accelerate")
@app.function(gpu="L4", image=image)
def run(prompt: str) -> str:
from transformers import pipeline
pipe = pipeline("text-generation", model="meta-llama/Llama-3.2-1B")
return pipe(prompt)[0]["generated_text"]
app.deploy()
print(run.remote("Hello from L4!"))Also available on Velar
Deploy on L4 24GB today
Start with $10 in free GPU credits. No credit card required. First workload live in under 60 seconds.