Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
35.4 tok/s
TTFT
5467 ms
Safe context
25K
Memory
30.5 GB / 48.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 40.3 tok/s | 6983 ms | 43K |
| Chat | C | Runs well | 40.3 tok/s | 2619 ms | 14K |
| Coding | C | Runs well | 35.4 tok/s | 5467 ms | 25K |
| RAG | C | Runs well | 40.3 tok/s | 8728 ms | 43K |
| Reasoning | C | Runs well | 40.3 tok/s | 5673 ms | 25K |
How Qwen 2.5 Coder 32B (32B params) fits at each quantization level on RTX 6000 Ada 48GB (48.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 12.5 GB | Low | D35 |
Q3_K_S | 3 | 15.7 GB | Low | D36 |
NVFP4 | 4 |
ollama run qwen-2.5-coder-32bhuggingface-cli download qwen-2.5-coder-32b17.9 GB |
| Medium |
| D37 |
Q4_K_M | 4 | 19.5 GB | Medium | D38 |
Q5_K_M | 5 | 23.0 GB | High | D40 |
Q6_K | 6 | 26.2 GB | High | C41 |
Q8_0Best for your GPU | 8 | 34.2 GB | Very High | C44 |
F16 | 16 | 65.6 GB | Maximum | F0 |