Loading…
Loading…
Models / Zamba2
3 models (1.2B to 7B) · ↓ 138.1K downloads a month · from $0.10/h on RTX 3060 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 |
|---|---|---|---|---|
| instructZyphra/Zamba2-1.2B-instruct | vLLM | 2.4 GB | 6.3 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalZyphra/Zamba2-2.7B | vLLM | 5.3 GB | 10.7 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| InstructZyphra/Zamba2-7B-Instruct | vLLM | 30 GB | 40.6 GB | $0.46/h RTX A6000 |
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").