✅ — Google Cloud A3 (8x H100) Can Run Mixtral 8x7B
The Google Cloud A3 (8x H100) packs 640GB VRAM. Mixtral 8x7B needs ~25.7GB. You're good to go.
✅ Plenty of headroom
Hardware Specs
- GPU
- Google Cloud A3 (8x H100)
- VRAM
- 640 GB
- Bandwidth
- 26400 GB/s
- MSRP
- $0
Model Requirements
- Model
- Mixtral 8x7B
- Parameters
- 46.7B
- VRAM (Q4)
- ~25.7 GB
- Context
- N/A
Estimated Performance
848 tok/s
Estimated inference speed (Q4 quantization)