Hardware /
Apple M3 Max (30-core GPU)
12measured results
133best gen tok/s*
36 GBreported memory
*Headline excludes tiny/test models (under 0.6B params or TinyStories-style). Open in rig checker →
llama.cpp
| Model | Quant | Gen tok/s | Prompt tok/s | Source | Context tested | Build | |
|---|---|---|---|---|---|---|---|
| Llama 3.2 1B Instruct | Q4_K - Medium | 133 | 2,553 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 3.2 1B Instruct | Q4_K - Medium | 133 | 2,520 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 2 7B | Q4_0 | 56.6 | 568 | llama.cpp (Metal) ↗ | 512+128 | llama.cpp Metal | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 37.7 | 457 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 37.7 | 442 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Meta Llama 3.1 8B Instruct | Q4_K - Medium | 37.3 | 445 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 2 7B | Q8_0 | 34.3 | 566 | llama.cpp (Metal) ↗ | 512+128 | llama.cpp Metal | Details → |
| Qwen2.5 14B Instruct | Q4_K - Medium | 20.9 | 242 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Qwen2.5 14B Instruct | Q4_K - Medium | 20.7 | 236 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Qwen2.5 14B Instruct | Q4_K - Medium | 20.5 | 230 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
| Llama 2 7B | F16 | 19.5 | 589 | llama.cpp (Metal) ↗ | 512+128 | llama.cpp Metal | Details → |
| Qwen2.5 14B Instruct | Q4_K - Medium | 19.0 | 222 | LocalScore ↗ | 512+128 | llamafile 0.9.2 | Details → |
Ollama
No measured Ollama results yet. Submit yours →
MLX
No measured MLX results yet. Submit yours →
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.