Loading…
Loading…
Models / Llama-3.2 / Llama-3.2 3B
Text generation · served with vLLM · ↓ 517.5K downloads a month · View on Hugging Face
The same model packaged differently: a quantized version needs less GPU memory and runs on a cheaper machine.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalmeta-llama/Llama-3.2-3B-InstructLicence to accept | vLLM | 6.4 GB | ~9.9 GB | $0.10/h RTX 3060 |
| FP8 dynamic (RedHatAI)RedHatAI/Llama-3.2-3B-Instruct-FP8-dynamic | vLLM | 4.4 GB | 8.2 GB | $0.10/h RTX 3060 |
| FP8 (RedHatAI)RedHatAI/Llama-3.2-3B-Instruct-FP8 | vLLM | 4.4 GB | 8.1 GB | $0.10/h RTX 3060 |
| unsloth bnb 4bit (unsloth)unsloth/Llama-3.2-3B-Instruct-unsloth-bnb-4bit | vLLM | 2.4 GB | 5.9 GB | $0.10/h RTX 3060 |
| bnb 4bit (unsloth)unsloth/Llama-3.2-3B-Instruct-bnb-4bit | vLLM | 2.2 GB | 5.8 GB | $0.10/h RTX 3060 |
| llama 3 2 3b instruct onnx genai int4 kquantlast emb int8 (Arm)Arm/llama-3-2-3b-instruct-onnx-genai-int4-kquantlast-emb-int8 | ONNX Runtime GenAI | 2.7 GB | 5.6 GB | $0.08/h GTX 1660 SUPER |
$0.68/h from your balance while it runs, billed by the hour; stop it from its page. GridPup rents the machine with vLLM, which downloads the model from Hugging Face (usually ready in 10 to 15 minutes), and gives you an https address, keys and a Playground.
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.