Q_pi_S
7 records across 1 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M4 Max | Q_pi_S | 6bit | oMLX | 39.9 | 201.5 | – | – | 32768 | 2026-10-09 | community | – | source |
| M4 Max | Q_pi_S | 6bit | oMLX | 30.1 | 176.1 | – | – | 65536 | 2026-10-09 | community | – | source |
| M4 Max | Q_pi_S | 6bit | oMLX | 25.7 | 136.3 | – | – | 131072 | 2026-10-08 | community | – | source |
| M4 Max | Q_pi_S | 6bit | oMLX | 25.6 | 123.8 | – | – | 131072 | 2026-10-08 | community | – | source |
| M4 Max | Q_pi_S | 6bit | oMLX | 25.4 | 138.9 | – | – | 131072 | 2026-10-09 | community | – | source |
| M4 Max | Q_pi_S | 6bit | oMLX | 23.4 | 137.4 | – | – | 131072 | 2026-10-08 | community | – | source |
| M4 Max | Q_pi_S | 6bit | oMLX | 20.2 | 123.1 | – | – | 131072 | 2026-10-08 | community | contradiction | source |