Apple M1 Max (24-core GPU) × Qwen2.5 14B Instruct
Qwen2.5 14B Instruct on Apple M1 Max (24-core GPU)
20.2best gen tok/s
160best prompt tok/s
4,352max 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 | 20.2 | 158 | LocalScore ↗ | 4,352 | 7,729 ms | llamafile 0.9.2 |
| Q4_K - Medium | 17.6 | 159 | LocalScore ↗ | 512+128 | 7,687 ms | llamafile 0.9.2 |
| Q4_K - Medium | 17.5 | 160 | LocalScore ↗ | 512+128 | 7,666 ms | llamafile 0.9.2 |
| Q4_K - Medium | 17.4 | 156 | LocalScore ↗ | 512+128 | 7,863 ms | llamafile 0.9.2 |
| Q4_K - Medium | 13.8 | 130 | LocalScore ↗ | 512+128 | 9,350 ms | llamafile 0.9.2 |
| Q4_K - Medium | 11.7 | 112 | LocalScore ↗ | 512+128 | 10,618 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 Max (24-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.