Can it run?
Runs well
Using Q4_K_M in vLLM
Fit status
Runs well
Decode
76.2 tok/s
TTFT
3001 ms
Safe context
37K
Memory
126.2 GB / 288.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 77.2 tok/s | 3647 ms | 63K |
| Chat | C | Runs well | 77.2 tok/s | 1368 ms | 20K |
| Coding | C | Runs well | 76.2 tok/s | 3001 ms | 37K |
| RAG | C | Runs well | 77.2 tok/s | 4559 ms | 63K |
| Reasoning | C | Runs well | 77.2 tok/s | 2963 ms | 37K |
How Pixtral Large 124B (124B params) fits at each quantization level on AMD Instinct MI350X 288GB (288.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 48.4 GB | Low | D33 |
Q3_K_S | 3 | 60.8 GB | Low | D34 |
NVFP4 | 4 |
huggingface-cli download pixtral-large-124b69.4 GB |
| Medium |
| D35 |
Q4_K_M | 4 | 75.6 GB | Medium | D35 |
Q5_K_M | 5 | 89.3 GB | High | D36 |
Q6_K | 6 | 101.7 GB | High | D37 |
Q8_0Best for your GPU | 8 | 132.7 GB | Very High | D39 |
F16 | 16 | 254.2 GB | Maximum | C44 |