Loading…
Loading…
Models / granite-3.1
4 models (1B to 8B) · ↓ 131.3K 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 |
|---|---|---|---|---|
| instructibm-granite/granite-3.1-1b-a400m-instruct | vLLM | 2.7 GB | 5.6 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| instructibm-granite/granite-3.1-2b-instruct | vLLM | 5.1 GB | 8.6 GB | $0.10/h RTX 3060 |
| granite 3.2 2b instructibm-granite/granite-3.2-2b-instruct | vLLM | 5.1 GB | 8.6 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| instructibm-granite/granite-3.1-3b-a800m-instruct | vLLM | 6.6 GB | 10.1 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| granite 3.2 8b instructibm-granite/granite-3.2-8b-instruct | vLLM | 16 GB | 21.9 GB | $0.18/h 2x RTX 3060 |
| instructibm-granite/granite-3.1-8b-instruct | vLLM | 16 GB | 21.9 GB | $0.18/h 2x 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").