M3 Max (30-core GPU, 36 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 Max (30-core GPU, 36 GB) | llama-3-1-8b-instruct | Q4_K - Medium | llamafile | 37.47979072222222 | 443.4985221666666 | – | 2.725718886444444 | – | – | sourced | – | source |
| M3 Max (30-core GPU, 36 GB) | llama-3-2-1b-instruct | Q4_K - Medium | llamafile | 132.9925948888889 | 2553.2959846666668 | – | 0.4669405787777778 | – | – | sourced | – | source |
| M3 Max (30-core GPU, 36 GB) | qwen-2-5-14b-instruct | Q4_K - Medium | llamafile | 19.768525888888888 | 226.180183 | – | 5.407046111222224 | – | – | sourced | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085