Loading…
Loading…
Models / Llama3
2 models (70B to 70B) · ↓ 17.2K downloads a month · from $1.78/h on 4x RTX A6000 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 |
|---|---|---|---|---|
| OriginalValiantLabs/Llama3-70B-ShiningValiant2 | vLLM | 141 GB | 162 GB | $1.78/h 4x RTX A6000 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalValiantLabs/Llama3-70B-Fireplace | vLLM | 141 GB | 162 GB | $1.78/h 4x 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").