Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
57.3 tok/s
TTFT
3378 ms
Safe context
21K
Memory
62.1 GB / 80.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Tight fit | 66.9 tok/s | 4212 ms | 35K |
| Chat | B | Runs well | 66.9 tok/s | 1579 ms | 11K |
| Coding | B | Runs well | 57.3 tok/s | 3378 ms | 21K |
| RAG | C | Tight fit | 66.9 tok/s | 5265 ms | 35K |
| Reasoning | B | Runs well | 66.9 tok/s | 3422 ms | 21K |
How Solar Open 69B REAP i1 (69B params) fits at each quantization level on NVIDIA H100 80GB (80.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 26.9 GB | Low | D36 |
Q3_K_S | 3 | 33.8 GB | Low | D38 |
NVFP4 | 4 |
38.6 GB |
| Medium |
| D39 |
Q4_K_M | 4 | 42.1 GB | Medium | C40 |
Q5_K_M | 5 | 49.7 GB | High | C42 |
Q6_KBest for your GPU | 6 | 56.6 GB | High | C44 |
Q8_0 | 8 | 73.8 GB | Very High | C44 |
F16 | 16 | 141.5 GB | Maximum | F0 |
