Apple M4 Pro (16-core GPU) × Qwen2.5 14B Instruct
Qwen2.5 14B Instruct on Apple M4 Pro (16-core GPU)
16.9best gen tok/s
161best prompt tok/s
4,352max context tested
11results
Measured · all numbers below come from the linked sources
llama.cpp
| Quant | Gen tok/s | Prompt tok/s | Source | Context tested | TTFT | Build |
|---|---|---|---|---|---|---|
| Q4_K - Medium | 16.9 | 156 | LocalScore ↗ | 512+128 | 8,130 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.8 | 161 | LocalScore ↗ | 512+128 | 7,844 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.7 | 154 | LocalScore ↗ | 4,352 | 8,239 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.7 | 157 | LocalScore ↗ | 4,352 | 8,025 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.7 | 153 | LocalScore ↗ | 512+128 | 8,316 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.6 | 153 | LocalScore ↗ | 4,352 | 8,281 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.6 | 153 | LocalScore ↗ | 512+128 | 8,254 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.5 | 150 | LocalScore ↗ | 512+128 | 8,487 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.5 | 156 | LocalScore ↗ | 512+128 | 8,101 ms | llamafile 0.9.2 |
| Q4_K - Medium | 16.1 | 151 | LocalScore ↗ | 512+128 | 8,461 ms | llamafile 0.9.2 |
| Q4_K - Medium | 10.9 | 116 | LocalScore ↗ | 512+128 | 10,946 ms | llamafile 0.9.2 |
Ollama
No measured Ollama results yet. Submit yours →
MLX
No measured MLX results yet. Submit yours →
"Context tested" is the largest prompt+generation length actually benchmarked (tokens), not a memory-fit estimate.
What else can Apple M4 Pro (16-core GPU) run? →
Includes results from LocalScore, a Mozilla Builders project by cjpais, mirrored in the cn0303/local-llm-speed-benchmarks dataset. LocalScore runs use llamafile, which is built on llama.cpp.