Can it run?
Runs well
Using Q4_K_M in llama.cpp
Fit status
Runs well
Decode
77.1 tok/s
TTFT
2512 ms
Safe context
21K
Memory
107.1 GB / 141.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | B | Runs well | 89.9 tok/s | 3131 ms | 40K |
| Chat | B | Runs well | 89.9 tok/s | 1174 ms | 11K |
| Coding | B | Runs well | 77.1 tok/s | 2512 ms | 21K |
| RAG | B | Runs well | 89.9 tok/s | 3914 ms | 40K |
| Reasoning | B | Runs well | 89.9 tok/s | 2544 ms | 21K |
How Mixtral 8x22B (141B params) fits at each quantization level on NVIDIA H200 141GB (141.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 55.0 GB | Low | D38 |
Q3_K_S | 3 | 69.1 GB | Low | C40 |
NVFP4 | 4 |
ollama run mixtral-8x22bhuggingface-cli download mixtral-8x22b79.0 GB |
| Medium |
| C42 |
Q4_K_M | 4 | 86.0 GB | Medium | C43 |
Q5_K_MBest for your GPU | 5 | 101.5 GB | High | C45 |
Q6_K | 6 | 115.6 GB | High | C45 |
Q8_0 | 8 | 150.9 GB | Very High | F0 |
F16 | 16 | 289.0 GB | Maximum | F0 |