Loading…
Loading…
Models / dolphin-2.5-mixtral-8x7b
GPU memory needed (smallest version of each): dolphin-2.5-mixtral-8x7b ~46.7B about 29.9 GB. The cheapest runs on RTX PRO 4500 from $0.40/h right now.
1 model (~46.7B) · ↓ 1.4K downloads a month · from $0.40/h on RTX PRO 4500 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 |
|---|---|---|---|---|
| Originaldphn/dolphin-2.5-mixtral-8x7b | vLLM | 93 GB | 107.2 GB | $1.44/h 8x RTX 3090 |
| GPTQ (TheBloke)TheBloke/dolphin-2.5-mixtral-8x7b-GPTQ | vLLM | 24 GB | 29.9 GB | $0.40/h RTX PRO 4500 |
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").