Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
141.6 tok/s
TTFT
1367 ms
Safe context
24K
Memory
4.1 GB / 6.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | B | Runs well | 102.3 tok/s | 2753 ms | 47K |
| Chat | B | Runs well | 102.3 tok/s | 1032 ms | 12K |
| Coding | B | Runs well | 141.6 tok/s | 1367 ms | 24K |
| RAG | B | Runs well | 102.3 tok/s | 3441 ms | 47K |
| Reasoning | B | Runs well | 102.3 tok/s | 2237 ms | 24K |
How EXAONE 3.5 2.4B Instruct (2.4000000953674316B params) fits at each quantization level on RX 5600 XT 6GB (6.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 0.9 GB | Low | D32 |
Q3_K_S | 3 | 1.2 GB | Low | D33 |
NVFP4 | 4 |
| Medium |
| D34 |
Q4_K_M | 4 | 1.5 GB | Medium | D34 |
Q5_K_M | 5 | 1.7 GB | High | D35 |
Q6_K | 6 | 2.0 GB | High | D36 |
Q8_0Best for your GPU | 8 | 2.6 GB | Very High | D38 |
F16 | 16 | 4.9 GB | Maximum | C44 |