Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
30.7 tok/s
TTFT
6309 ms
Safe context
23K
Memory
33.6 GB / 48.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 30.7 tok/s | 9176 ms | 39K |
| Chat | C | Runs well | 30.7 tok/s | 3441 ms | 12K |
| Coding | C | Runs well | 30.7 tok/s | 6309 ms | 23K |
| RAG | C | Runs well | 30.7 tok/s | 11470 ms | 39K |
| Reasoning | C | Runs well | 30.7 tok/s | 7456 ms | 23K |
How Hermes 4.3 36B (36B params) fits at each quantization level on NVIDIA L40 48GB (48.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 14.0 GB | Low | D35 |
Q3_K_S | 3 | 17.6 GB | Low | D37 |
NVFP4 | 4 | 20.2 GB | Medium | D38 |
Q4_K_M | 4 | 22.0 GB | Medium | D39 |
Q5_K_M | 5 | 25.9 GB | High | C41 |
Q6_K | 6 | 29.5 GB | High | C42 |
Q8_0Best for your GPU | 8 | 38.5 GB | Very High | C44 |
F16 | 16 | 73.8 GB | Maximum | F0 |
