Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
201.6 tok/s
TTFT
960 ms
Safe context
34K
Memory
3.7 GB / 8.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 142.8 tok/s | 1972 ms | 69K |
| Chat | C | Runs well | 142.8 tok/s | 739 ms | 17K |
| Coding | C | Runs well | 201.6 tok/s | 960 ms | 34K |
| RAG | C | Runs well | 142.8 tok/s | 2465 ms | 69K |
| Reasoning | C | Runs well | 142.8 tok/s | 1602 ms | 34K |
How Qwen2.5 1.5B Instruct (1.5B params) fits at each quantization level on RX 6650 XT 8GB (8.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 0.6 GB | Low | D31 |
Q3_K_S | 3 | 0.7 GB | Low | D31 |
NVFP4 | 4 |
| Medium |
| D32 |
Q4_K_M | 4 | 0.9 GB | Medium | D32 |
Q5_K_M | 5 | 1.1 GB | High | D32 |
Q6_K | 6 | 1.2 GB | High | D33 |
Q8_0 | 8 | 1.6 GB | Very High | D34 |
F16Best for your GPU | 16 | 3.1 GB | Maximum | D38 |