Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
40.9 tok/s
TTFT
4731 ms
Safe context
29K
Memory
26.7 GB / 48.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 40.9 tok/s | 6882 ms | 50K |
| Chat | C | Runs well | 40.9 tok/s | 2581 ms | 16K |
| Coding | C | Runs well | 40.9 tok/s | 4731 ms | 29K |
| RAG | C | Runs well | 40.9 tok/s | 8603 ms | 50K |
| Reasoning | C | Runs well | 40.9 tok/s | 5592 ms | 29K |
How Qwen3.5 27B (27B params) fits at each quantization level on NVIDIA L40S 48GB (48.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 10.5 GB | Low | D34 |
Q3_K_S | 3 | 13.2 GB | Low | D35 |
NVFP4 | 4 | 15.1 GB | Medium | D36 |
Q4_K_M | 4 | 16.5 GB | Medium | D37 |
Q5_K_M | 5 | 19.4 GB | High | D38 |
Q6_K | 6 | 22.1 GB | High | D39 |
Q8_0Best for your GPU | 8 | 28.9 GB | Very High | C42 |
F16 | 16 | 55.4 GB | Maximum | F0 |
huggingface-cli download hf-unsloth--qwen3-5-27b-gguf