Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
51.0 tok/s
TTFT
3797 ms
Safe context
20K
Memory
32.8 GB / 40.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs with offload | 59.5 tok/s | 4734 ms | 33K |
| Chat | B | Runs well | 59.5 tok/s | 1775 ms | 11K |
| Coding | C | Runs well | 51.0 tok/s | 3797 ms | 20K |
| RAG | C | Runs with offload | 59.5 tok/s | 5918 ms | 33K |
| Reasoning | B | Runs well | 59.5 tok/s | 3847 ms | 20K |
How Hermes 4.3 36B (36B params) fits at each quantization level on NVIDIA A100 40GB (40.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 14.0 GB | Low | D36 |
Q3_K_S | 3 | 17.6 GB | Low | D38 |
NVFP4 | 4 |
20.2 GB |
| Medium |
| D40 |
Q4_K_M | 4 | 22.0 GB | Medium | C41 |
Q5_K_M | 5 | 25.9 GB | High | C43 |
Q6_KBest for your GPU | 6 | 29.5 GB | High | C44 |
Q8_0 | 8 | 38.5 GB | Very High | C44 |
F16 | 16 | 73.8 GB | Maximum | F0 |