M1 Max (32-core GPU, 64 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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M1 Max (32-core GPU, 64 GB) | llama-3-1-8b-instruct | Q4_K - Medium | llamafile | 37.75919747222222 | 380.7280225277778 | – | 3.12193795138889 | – | – | sourced | – | source |
| M1 Max (32-core GPU, 64 GB) | llama-3-2-1b-instruct | Q4_K - Medium | llamafile | 120.67175636111112 | 1972.074889361111 | – | 0.5782512951666667 | – | – | sourced | – | source |
| M1 Max (32-core GPU, 64 GB) | qwen-2-5-14b-instruct | Q4_K - Medium | llamafile | 19.01644427777778 | 185.93018677777775 | – | 6.8818776490777775 | – | – | sourced | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085