Loading…
Loading…
Models / ministral-3
GPU memory needed (smallest version of each): ministral-3 3B about 4.7 GB, ministral-3 8B about 9.1 GB. The cheapest runs on Quadro P4000 from $0.07/h right now.
2 models (3B to 8B) · ↓ 0 downloads a month · from $0.07/h on Quadro P4000 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| onnxhyperquader/ministral-3-3b-amberlin-tools-onnx | ONNX Runtime GenAI | 2.0 GB | 4.7 GB | $0.07/h Quadro P4000 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| onnxhyperquader/ministral-3-8b-amberlin-tools-onnx | ONNX Runtime GenAI | 5.6 GB | 9.1 GB | $0.09/h RTX 3060 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").