gemma-4-12B-it-qat
41 records across 11 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M5 Ultra | gemma-4-12B-it-qat | 4bit | oMLX | 71.6 | 882.3 | – | – | 1024 | 2026-10-16 | community | – | source |
| M5 Ultra | gemma-4-12B-it-qat | 4bit | oMLX | 70.6 | 2171.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M5 Ultra | gemma-4-12B-it-qat | 4bit | oMLX | 69.1 | 2399.0 | – | – | 16384 | 2026-10-16 | community | – | source |
| M3 Ultra | gemma-4-12B-it-qat | oQ4e | oMLX | 65.9 | 645.6 | – | – | 1024 | 2026-10-10 | community | – | source |
| M3 Ultra | gemma-4-12B-it-qat | oQ4e | oMLX | 65.1 | 693.0 | – | – | 4096 | 2026-10-10 | community | – | source |
| M3 Ultra | gemma-4-12B-it-qat | oQ4e | oMLX | 64.5 | 676.4 | – | – | 8192 | 2026-10-10 | community | – | source |
| M3 Ultra | gemma-4-12B-it-qat | oQ4e | oMLX | 62.8 | 648.2 | – | – | 16384 | 2026-10-10 | community | – | source |
| M5 Ultra | gemma-4-12B-it-qat | 4bit | oMLX | 59.2 | 1602.0 | – | – | 131072 | 2026-10-16 | community | – | source |
| M4 Pro | gemma-4-12B-it-qat | oQ4e | oMLX | 45.4 | 224.3 | – | – | 1024 | 2026-10-10 | community | – | source |
| M5 Max | gemma-4-12B-it-qat | 4bit | oMLX | 44.1 | 1704.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| M5 Max | gemma-4-12B-it-qat | 4bit | oMLX | 40.5 | 1278.0 | – | – | 65536 | 2026-10-16 | community | – | source |
| M4 Pro | gemma-4-12B-it-qat | oQ4e | oMLX | 37.2 | 216.0 | – | – | 8192 | 2026-10-10 | community | – | source |
| M5 Max | gemma-4-12B-it-qat | 4bit | oMLX | 36.5 | 1002.0 | – | – | 131072 | 2026-10-16 | community | – | source |
| M5 Max | gemma-4-12B-it-qat | 8bit | oMLX | 36.3 | 1096.0 | – | – | 1024 | 2026-10-13 | community | – | source |
| M5 Max | gemma-4-12B-it-qat | 8bit | oMLX | 35.5 | 1517.0 | – | – | 4096 | 2026-10-13 | community | – | source |
| M4 Pro | gemma-4-12B-it-qat | oQ4e | oMLX | 34.9 | 218.2 | – | – | 4096 | 2026-10-10 | community | – | source |
| M5 Max | gemma-4-12B-it-qat | 4bit | oMLX | 33.4 | 809.9 | – | – | 199680 | 2026-10-16 | community | – | source |
| M4 Pro | gemma-4-12B-it-qat | oQ4e | oMLX | 29.3 | 220.5 | – | – | 4096 | 2026-10-10 | community | contradiction | source |
| M4 Pro | gemma-4-12B-it-qat | oQ4e | oMLX | 27.7 | 209.4 | – | – | 1024 | 2026-10-10 | community | contradiction | source |
| M1 Max | gemma-4-12B-it-qat | 4bit | oMLX | 25.3 | 230.4 | – | – | 1024 | 2026-10-17 | community | – | source |
| M1 Max | gemma-4-12B-it-qat | 4bit | oMLX | 25.0 | 238.2 | – | – | 4096 | 2026-10-17 | community | – | source |
| M5 Pro | gemma-4-12B-it-qat | 4bit | oMLX | 22.9 | 825.9 | – | – | 32768 | 2026-10-17 | community | – | source |
| M4 | gemma-4-12B-it-qat | oQ4e | oMLX | 22.1 | 136.0 | – | – | 1024 | 2026-10-13 | community | – | source |
| M5 Pro | gemma-4-12B-it-qat | 4bit | oMLX | 21.9 | 714.0 | – | – | 65536 | 2026-10-17 | community | – | source |
| M5 Pro | gemma-4-12B-it-qat | 4bit | oMLX | 19.9 | 577.4 | – | – | 131072 | 2026-10-17 | community | – | source |
| M6 | gemma-4-12B-it-qat | oQ4e | oMLX | 19.6 | 560.9 | – | – | 1024 | 2026-10-08 | community | – | source |
| M6 | gemma-4-12B-it-qat | oQ4e | oMLX | 19.3 | 625.4 | – | – | 4096 | 2026-10-08 | community | – | source |
| M5 Pro | gemma-4-12B-it-qat | 4bit | oMLX | 18.5 | 463.9 | – | – | 199680 | 2026-10-17 | community | – | source |
| M3 Pro | gemma-4-12B-it-qat | 8bit | oMLX | 17.2 | 213.1 | – | – | 1024 | 2026-10-12 | community | – | source |
| M2 Pro | gemma-4-12B-it-qat | 4bit | oMLX | 16.3 | 127.4 | – | – | 4096 | 2026-10-10 | community | – | source |
| M2 Pro | gemma-4-12B-it-qat | 4bit | oMLX | 16.0 | 124.1 | – | – | 16384 | 2026-10-10 | community | – | source |
| M3 Pro | gemma-4-12B-it-qat | 8bit | oMLX | 14.2 | 202.1 | – | – | 4096 | 2026-10-12 | community | – | source |
| M4 | gemma-4-12B-it-qat | oQ4e | oMLX | 13.3 | 137.4 | – | – | 4096 | 2026-10-13 | community | – | source |
| M1 Max | gemma-4-12B-it-qat | bf16 | oMLX | 12.9 | 297.4 | – | – | 1024 | 2026-10-12 | community | – | source |
| M1 Max | gemma-4-12B-it-qat | bf16 | oMLX | 12.8 | 307.9 | – | – | 4096 | 2026-10-12 | community | – | source |
| M3 Pro | gemma-4-12B-it-qat | 8bit | oMLX | 10.2 | 211.4 | – | – | 1024 | 2026-10-12 | community | contradiction | source |
| M3 Pro | gemma-4-12B-it-qat | 8bit | oMLX | 9.9 | 208.5 | – | – | 4096 | 2026-10-12 | community | contradiction | source |
| M4 | gemma-4-12B-it-qat | oQ4e | oMLX | 9.3 | 75.5 | – | – | 4096 | 2026-10-08 | community | contradiction | source |
| AMD Custom GPU 0932 | gemma-4-12B-it-qat 11.91B | Q4_0 | llama.cpp (Vulkan) | 9.23 | 90.83 | – | – | tg128 | 2026-08-20 | community | – | source |
| M4 | gemma-4-12B-it-qat | oQ4e | oMLX | 9.2 | 76.5 | – | – | 1024 | 2026-10-08 | community | contradiction | source |
| AMD Custom GPU 0932 | gemma-4-12B-it-qat 11.91B | Q4_0 | llama.cpp (Vulkan) | 9.06 | 84.66 | – | – | tg128 | 2026-08-20 | community | – | source |