RTX A6000 48GBBudget pickC29.9 tok/s decode
~$4,650 MSRP
Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
11.3 tok/s
TTFT
17178 ms
Safe context
23K
Memory
32.6 GB / 46.1 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 11.3 tok/s | 24986 ms | 39K |
| Chat | C | Runs well | 11.3 tok/s | 9370 ms | 12K |
| Coding | C | Runs well | 11.3 tok/s | 17178 ms | 23K |
| RAG | C | Runs well | 11.3 tok/s | 31232 ms | 39K |
| Reasoning | C | Runs well | 11.3 tok/s | 20301 ms | 23K |
How Qwen 2.5 Coder 32B (32B params) fits at each quantization level on MacBook Pro M1 Max 64GB (46.1 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 12.5 GB | Low | D35 |
Q3_K_S | 3 | 15.7 GB | Low | D37 |
NVFP4 | 4 | 17.9 GB | Medium | D38 |
Q4_K_M | 4 | 19.5 GB | Medium | D38 |
Q5_K_M | 5 | 23.0 GB | High | C40 |
Q6_K | 6 | 26.2 GB | High | C42 |
Q8_0Best for your GPU | 8 | 34.2 GB | Very High | C44 |
F16 | 16 | 65.6 GB | Maximum | F0 |
ollama run qwen-2.5-coder-32bhuggingface-cli download qwen-2.5-coder-32bUpgrade options
RTX A6000 48GBBudget pick~$4,650 MSRP
NVIDIA L40 48GBBest value~$5,500 MSRP
RTX PRO 5000 Blackwell 48GBBiggest leap