gemma-4-31B-it-qat
30 records across 4 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M5 Max | gemma-4-31B-it-qat | 4bit | oMLX | 21.8 | 694.8 | – | – | 1024 | 2026-10-10 | community | – | source |
| M5 Max | gemma-4-31B-it-qat | 4bit | oMLX | 21.1 | 708.1 | – | – | 4096 | 2026-10-10 | community | – | source |
| M5 Max | gemma-4-31B-it-qat | 4bit | oMLX | 19.8 | 638.7 | – | – | 8192 | 2026-10-10 | community | – | source |
| M5 Max | gemma-4-31B-it-qat | 4bit | oMLX | 18.3 | 556.4 | – | – | 16384 | 2026-10-10 | community | – | source |
| M5 Max | gemma-4-31B-it-qat | 4bit | oMLX | 16.8 | 620.6 | – | – | 32768 | 2026-10-11 | community | – | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 16.5 | 90.7 | – | – | 1024 | 2026-10-17 | community | – | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 15.6 | 89.8 | – | – | 4096 | 2026-10-17 | community | – | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 15.0 | 90.1 | – | – | 1024 | 2026-10-17 | community | – | source |
| M5 Max | gemma-4-31B-it-qat | 4bit | oMLX | 13.5 | 519.8 | – | – | 65536 | 2026-10-11 | community | – | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 13.1 | 89.8 | – | – | 4096 | 2026-10-17 | community | contradiction | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 12.9 | 89.8 | – | – | 1024 | 2026-10-17 | community | contradiction | source |
| M4 Max | gemma-4-31B-it-qat | 8bit | oMLX | 12.7 | 175.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Max | gemma-4-31B-it-qat | oQ8 | oMLX | 12.7 | 175.8 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Max | gemma-4-31B-it-qat | 8bit | oMLX | 12.7 | 175.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Max | gemma-4-31B-it-qat | oQ8 | oMLX | 12.7 | 175.8 | – | – | 16384 | 2026-10-11 | community | – | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 12.5 | 89.7 | – | – | 4096 | 2026-10-17 | community | contradiction | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 12.4 | 90.7 | – | – | 1024 | 2026-10-17 | community | contradiction | source |
| M4 Max | gemma-4-31B-it-qat | oQ8 | oMLX | 12.0 | 169.7 | – | – | 32768 | 2026-10-11 | community | – | source |
| M4 Max | gemma-4-31B-it-qat | oQ8 | oMLX | 12.0 | 169.7 | – | – | 32768 | 2026-10-11 | community | – | source |
| M4 Max | gemma-4-31B-it-qat | 8bit | oMLX | 11.9 | 168.9 | – | – | 32768 | 2026-10-11 | community | – | source |
| M4 Max | gemma-4-31B-it-qat | 8bit | oMLX | 11.9 | 168.9 | – | – | 32768 | 2026-10-11 | community | – | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 11.9 | 90.6 | – | – | 1024 | 2026-10-17 | community | contradiction | source |
| M5 Pro | gemma-4-31B-it-qat | 4bit | oMLX | 11.6 | 365.0 | – | – | 1024 | 2026-10-14 | community | – | source |
| M5 Pro | gemma-4-31B-it-qat | 4bit | oMLX | 11.6 | 365.0 | – | – | 1024 | 2026-10-14 | community | – | source |
| M5 Pro | gemma-4-31B-it-qat | 4bit | oMLX | 11.3 | 359.8 | – | – | 4096 | 2026-10-14 | community | – | source |
| M5 Pro | gemma-4-31B-it-qat | 4bit | oMLX | 11.3 | 359.8 | – | – | 4096 | 2026-10-14 | community | – | source |
| M5 Pro | gemma-4-31B-it-qat | 4bit | oMLX | 10.2 | 316.5 | – | – | 16384 | 2026-10-14 | community | – | source |
| M5 Pro | gemma-4-31B-it-qat | 4bit | oMLX | 10.2 | 316.5 | – | – | 16384 | 2026-10-14 | community | – | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 10.2 | 88.8 | – | – | 4096 | 2026-10-17 | community | contradiction | source |
| M1 Max | gemma-4-31B-it-qat | 4bit | oMLX | 10.0 | 89.0 | – | – | 4096 | 2026-10-17 | community | contradiction | source |