Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
90.9 tok/s
TTFT
2129 ms
Safe context
32K
Memory
10.1 GB / 20.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 90.9 tok/s | 3097 ms | 56K |
| Chat | C | Runs well | 90.9 tok/s | 1161 ms | 17K |
| Coding | C | Runs well | 90.9 tok/s | 2129 ms | 32K |
| RAG | C | Runs well | 90.9 tok/s | 3871 ms | 56K |
| Reasoning | C | Runs well | 90.9 tok/s | 2516 ms | 32K |
How Qwen3.5 9B (9B params) fits at each quantization level on RTX A4500 20GB (20.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 3.5 GB | Low | D33 |
Q3_K_S | 3 | 4.4 GB | Low | D34 |
NVFP4 | 4 | 5.0 GB | Medium | D35 |
Q4_K_M | 4 | 5.5 GB | Medium | D35 |
Q5_K_M | 5 | 6.5 GB | High | D37 |
Q6_K | 6 | 7.4 GB | High | D37 |
Q8_0Best for your GPU | 8 | 9.6 GB | Very High | D40 |
F16 | 16 | 18.5 GB | Maximum | C45 |
huggingface-cli download hf-unsloth--qwen3-5-9b-gguf