Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
78.3 tok/s
TTFT
2471 ms
Safe context
8K
Memory
25.1 GB / 32.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Tight fit | 78.3 tok/s | 3595 ms | 8K |
| Chat | B | Runs well | 78.3 tok/s | 1348 ms | 8K |
| Coding | B | Runs well | 78.3 tok/s | 2471 ms | 8K |
| RAG | C | Tight fit | 78.3 tok/s | 4493 ms | 8K |
| Reasoning | B | Runs well | 78.3 tok/s | 2921 ms | 8K |
How Gemma 2 27B (27B params) fits at each quantization level on RTX 5090 32GB (32.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 10.5 GB | Low | D36 |
Q3_K_S | 3 | 13.2 GB | Low | D38 |
NVFP4 | 4 |
ollama run gemma-2-27bhuggingface-cli download gemma-2-27b| Medium |
| D39 |
Q4_K_M | 4 | 16.5 GB | Medium | C40 |
Q5_K_M | 5 | 19.4 GB | High | C42 |
Q6_KBest for your GPU | 6 | 22.1 GB | High | C44 |
Q8_0 | 8 | 28.9 GB | Very High | C44 |
F16 | 16 | 55.4 GB | Maximum | F0 |