Can it run?
Runs well
Using Q6_K in Ollama
Fit status
Runs well
Decode
136.5 tok/s
TTFT
1418 ms
Safe context
41K
Memory
4.5 GB / 11.5 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 136.5 tok/s | 2063 ms | 81K |
| Chat | C | Runs well | 136.5 tok/s | 774 ms | 20K |
| Coding | C | Runs well | 136.5 tok/s | 1418 ms | 41K |
| RAG | C | Runs well | 136.5 tok/s | 2579 ms | 81K |
| Reasoning | C | Runs well | 136.5 tok/s | 1676 ms | 41K |
How Llama 3.2 1B Instruct Q8 0 (1B params) fits at each quantization level on MacBook Pro M1 Pro 16GB (11.5 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 0.4 GB | Low | D30 |
Q3_K_S | 3 | 0.5 GB | Low | D30 |
NVFP4 | 4 | 0.6 GB | Medium | D31 |
Q4_K_M | 4 | 0.6 GB | Medium | D31 |
Q5_K_M | 5 | 0.7 GB | High | D31 |
Q6_K | 6 | 0.8 GB | High | D31 |
Q8_0 | 8 | 1.1 GB | Very High | D32 |
F16Best for your GPU | 16 | 2.1 GB | Maximum | D33 |
huggingface-cli download hf-hugging-quants--llama-3-2-1b-instruct-q8-0-gguf