Text generation · served with ONNX Runtime GenAI · ↓ 27 downloads a month · View on Hugging Face
inference4j/phi-4-mini-instruct needs about 8.2 GB of GPU memory. It runs on one RTX 3060 from $0.08/h right now; best value: RTX 5090 at $0.67/h.
Weights
4.9 GB
GPU memory needed
about 8.2 GB
Memory · disk
11.4 GB · 41 GB
Licence
mit
Hardware and deploy
Deploying ONNX models from the catalog is coming soon. Today, download the files from Hugging Face and upload them in Inference.
Prices are per hour for the whole machine and its disk, right now. Speed is Vast's DLPerf score of the machine, relative to the cheapest. Best value: the most speed per dollar. Only machines with a Vast reliability above 95% that download at 400 Mbit/s or more are offered: the machine fetches the model itself.
Questions
How much GPU memory does phi-4-mini-instruct need?
About 8.2 GB: 4.93 GB of weights and the runtime's own memory.
What is the cheapest GPU to run phi-4-mini-instruct?
Right now one RTX 3060 at $0.08 an hour; RTX 5090 at $0.67 an hour is the best value for its speed. Prices are live offers from GPU clouds, checked every 15 minutes.
Can phi-4-mini-instruct run on an RTX 4090?
Yes: it needs about 8.2 GB and an RTX 4090 has 24 GB.