Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
68.3 tok/s
TTFT
2833 ms
Safe context
31K
Memory
8.2 GB / 16.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 68.3 tok/s | 4120 ms | 55K |
| Chat | C | Runs well | 68.3 tok/s | 1545 ms | 16K |
| Coding | C | Runs well | 68.3 tok/s | 2833 ms | 31K |
| RAG | C | Runs well | 68.3 tok/s | 5150 ms | 55K |
| Reasoning | C | Runs well | 68.3 tok/s | 3348 ms | 31K |
How Llama 2 7B Chat (7B params) fits at each quantization level on RX 6900 XT 16GB (16.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 2.7 GB | Low | D33 |
Q3_K_S | 3 | 3.4 GB | Low | D34 |
NVFP4 | 4 | 3.9 GB | Medium | D35 |
Q4_K_M | 4 | 4.3 GB | Medium | D35 |
Q5_K_M | 5 | 5.0 GB | High | D36 |
Q6_K | 6 | 5.7 GB | High | D37 |
Q8_0Best for your GPU | 8 | 7.5 GB | Very High | D40 |
F16 | 16 | 14.3 GB | Maximum | C44 |
