M3 Pro (18-core GPU, 18 GB)
3 records across 3 models (quants: Q4_K - Medium).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M3 Pro (18-core GPU, 18 GB) | llama-3-1-8b-instruct | Q4_K - Medium | llamafile | 20.80849266666667 | 279.21399711111115 | – | 4.529956342555556 | – | – | sourced | – | source |
| M3 Pro (18-core GPU, 18 GB) | llama-3-2-1b-instruct | Q4_K - Medium | llamafile | 85.63227766666667 | 1573.6998116666664 | – | 0.8258373006666666 | – | – | sourced | – | source |
| M3 Pro (18-core GPU, 18 GB) | qwen-2-5-14b-instruct | Q4_K - Medium | llamafile | 11.63579448148148 | 144.77590885185188 | – | 8.883480046185184 | – | – | sourced | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085