Can it run?
Runs well
Using Q5_K_M in llama.cpp
Fit status
Runs well
Decode
73.7 tok/s
TTFT
2628 ms
Safe context
4K
Memory
13.0 GB / 16.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Tight fit | 73.7 tok/s | 3823 ms | 4K |
| Chat | B | Runs well | 73.7 tok/s | 1433 ms | 4K |
| Coding | B | Runs well | 73.7 tok/s | 2628 ms | 4K |
| RAG | C | Tight fit | 73.7 tok/s | 4778 ms | 4K |
| Reasoning | B | Runs well | 73.7 tok/s | 3106 ms | 4K |
How StableLM 2 12B (12B params) fits at each quantization level on RTX 5080 16GB (16.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 4.7 GB | Low | D35 |
Q3_K_S | 3 | 5.9 GB | Low | D36 |
NVFP4 | 4 |
huggingface-cli download stablelm-2-12b| Medium |
| D37 |
Q4_K_M | 4 | 7.3 GB | Medium | D38 |
Q5_K_M | 5 | 8.6 GB | High | D40 |
Q6_KBest for your GPU | 6 | 9.8 GB | High | C42 |
Q8_0 | 8 | 12.8 GB | Very High | C43 |
F16 | 16 | 24.6 GB | Maximum | F0 |