Loading…
Loading…
Models / calme-3.2-instruct
GPU memory needed (smallest version of each): calme-3.2-instruct 78B about 178.7 GB. The cheapest runs on 4x Nvidia L40S from $3.52/h right now.
1 model (78B) · ↓ 330 downloads a month · from $3.52/h on 4x Nvidia L40S 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 |
|---|---|---|---|---|
| OriginalMaziyarPanahi/calme-3.2-instruct-78b | vLLM | 156 GB | 178.7 GB | $3.52/h 4x Nvidia L40S |
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").