Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
600.7 tok/s
TTFT
350 ms
Safe context
89K
Memory
46.2 GB / 256.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 600.7 tok/s | 469 ms | 176K |
| Chat | C | Runs well | 600.7 tok/s | 350 ms | 44K |
| Coding | C | Runs well | 600.7 tok/s | 350 ms | 89K |
| RAG | C | Runs well | 600.7 tok/s | 586 ms | 176K |
| Reasoning | C | Runs well | 600.7 tok/s | 381 ms | 89K |
How Qwen3-Coder 30B A3B Instruct (30.5B params) fits at each quantization level on AMD Instinct MI325X 256GB (256.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 11.9 GB | Low | D31 |
Q3_K_S | 3 | 14.9 GB | Low | D31 |
NVFP4 | 4 |
ollama run qwen-3-coder-30b-a3bhuggingface-cli download qwen-3-coder-30b-a3b17.1 GB |
| Medium |
| D31 |
Q4_K_M | 4 | 18.6 GB | Medium | D31 |
Q5_K_M | 5 | 22.0 GB | High | D31 |
Q6_K | 6 | 25.0 GB | High | D32 |
Q8_0 | 8 | 32.6 GB | Very High | D32 |
F16Best for your GPU | 16 | 62.5 GB | Maximum | D35 |