Deploy open LLMs on GPUs by the hour.

Pick a model and a card. We rent it on vast.ai, start vLLM and give you an endpoint.

Deploy a model See pricing
A cat asleep belly-up on a graphics card
Starter Llama 3.1 8B Instruct · GeForce RTX 4090 1× 24 GB · fp16 $0.35/hr Pro Qwen3 30B-A3B Instruct 2507 · A100 80GB 1× 80 GB · fp16 $1.27/hr Performance Llama 3.3 70B Instruct · H100 SXM 80GB 4× 80 GB · fp16 $7.60/hr

All 190 configurations →