Apple M2 Pro (16-core GPU) × Meta Llama 3.1 8B Instruct
Meta Llama 3.1 8B Instruct on Apple M2 Pro (16-core GPU)
25.0best gen tok/s
229best prompt tok/s
4,352max context tested
5results
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 | 25.0 | 229 | LocalScore ↗ | 512+128 | 5,430 ms | llamafile 0.9.2 |
| Q4_K - Medium | 24.5 | 224 | LocalScore ↗ | 4,352 | 5,677 ms | llamafile 0.9.2 |
| Q4_K - Medium | 24.3 | 226 | LocalScore ↗ | 512+128 | 5,560 ms | llamafile 0.9.2 |
| Q4_K - Medium | 23.8 | 209 | LocalScore ↗ | 512+128 | 6,068 ms | llamafile 0.9.2 |
| Q4_K - Medium | 23.6 | 219 | LocalScore ↗ | 512+128 | 5,733 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 M2 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.