Apple M1 Pro (14-core GPU) × Qwen2.5 14B Instruct
Qwen2.5 14B Instruct on Apple M1 Pro (14-core GPU)
11.0best gen tok/s
94.8best prompt tok/s
640max context tested
6results
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 | 11.0 | 94.8 | LocalScore ↗ | 512+128 | 13,243 ms | llamafile 0.9.2 |
| Q4_K - Medium | 10.8 | 93.9 | LocalScore ↗ | 512+128 | 13,391 ms | llamafile 0.9.2 |
| Q4_K - Medium | 10.8 | 93.2 | LocalScore ↗ | 512+128 | 13,367 ms | llamafile 0.9.2 |
| Q4_K - Medium | 10.8 | 92.5 | LocalScore ↗ | 512+128 | 13,389 ms | llamafile 0.9.2 |
| Q4_K - Medium | 10.8 | 88.0 | LocalScore ↗ | 512+128 | 14,050 ms | llamafile 0.9.2 |
| Q4_K - Medium | 10.4 | 88.9 | LocalScore ↗ | 512+128 | 14,166 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 M1 Pro (14-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.