Mellum2.1-12B-A2.5B-Thinking
102 records across 9 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 134.5 | 4163.0 | – | – | 1024 | 2026-10-09 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 131.9 | 1574.0 | – | – | 1024 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 131.6 | 4153.0 | – | – | 1024 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 129.8 | 4044.0 | – | – | 4096 | 2026-10-09 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 125.6 | 2292.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 125.1 | 3827.0 | – | – | 4096 | 2026-10-09 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 124.6 | 2287.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 123.1 | 2200.0 | – | – | 8192 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 122.4 | 2199.0 | – | – | 8192 | 2026-10-11 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 119.5 | 1253.0 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 119.5 | 1253.0 | – | – | 1024 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 118.8 | 2054.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 118.8 | 2054.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 114.9 | 1182.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 114.9 | 1182.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 113.6 | 2085.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 112.8 | 1723.0 | – | – | 4096 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ5e | oMLX | 112.8 | 3638.0 | – | – | 4096 | 2026-10-15 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 111.6 | 1682.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 111.3 | 2744.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 109.8 | 3143.0 | – | – | 1024 | 2026-10-09 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 109.6 | 1788.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 109.5 | 2248.0 | – | – | 1024 | 2026-10-14 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ4e | oMLX | 109.0 | 1211.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ4e | oMLX | 109.0 | 1211.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 109.0 | 1679.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 109.0 | 1679.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 108.7 | 2606.0 | – | – | 4096 | 2026-10-14 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ5e | oMLX | 108.0 | 3653.0 | – | – | 8192 | 2026-10-15 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 105.7 | 1296.0 | – | – | 1024 | 2026-10-15 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 105.5 | 1537.0 | – | – | 16384 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ5e | oMLX | 105.0 | 3463.0 | – | – | 16384 | 2026-10-15 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 105.0 | 1097.0 | – | – | 1024 | 2026-10-15 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 104.2 | 3301.0 | – | – | 8192 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 103.5 | 3410.0 | – | – | 4096 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 102.6 | 3165.0 | – | – | 1024 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 102.2 | 3450.0 | – | – | 16384 | 2026-10-16 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 101.4 | 3167.0 | – | – | 16384 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 100.8 | 3429.0 | – | – | 4096 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 100.4 | 3333.0 | – | – | 8192 | 2026-10-09 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ4e | oMLX | 100.0 | 1087.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ4e | oMLX | 100.0 | 1087.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 99.5 | 3190.0 | – | – | 16384 | 2026-10-14 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 98.9 | 1781.0 | – | – | 1024 | 2026-10-10 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 96.7 | 1571.0 | – | – | 4096 | 2026-10-10 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 8bit | oMLX | 94.7 | 1361.0 | – | – | 65536 | 2026-10-11 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 93.8 | 1647.0 | – | – | 8192 | 2026-10-10 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 92.0 | 2873.0 | – | – | 32768 | 2026-10-09 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 91.9 | 1037.0 | – | – | 1024 | 2026-10-17 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 91.9 | 1037.0 | – | – | 1024 | 2026-10-17 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 89.9 | 1544.0 | – | – | 16384 | 2026-10-10 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 89.7 | 875.0 | – | – | 1024 | 2026-10-14 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | oQ4e | oMLX | 89.2 | 1019.0 | – | – | 1024 | 2026-10-13 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 87.9 | 1127.0 | – | – | 4096 | 2026-10-15 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | oQ4e | oMLX | 86.7 | 920.6 | – | – | 4096 | 2026-10-13 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 86.7 | 835.3 | – | – | 4096 | 2026-10-14 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 85.7 | 1028.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 85.7 | 1028.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 83.6 | 730.0 | – | – | 1024 | 2026-10-09 | community | contradiction | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 83.4 | 2228.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 82.6 | 2192.0 | – | – | 8192 | 2026-10-11 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 82.3 | 1358.0 | – | – | 32768 | 2026-10-10 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 81.6 | 997.4 | – | – | 4096 | 2026-10-15 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 81.1 | 2048.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 81.1 | 2048.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 80.9 | 2050.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 80.9 | 2050.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 80.9 | 837.8 | – | – | 16384 | 2026-10-17 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 80.9 | 837.8 | – | – | 16384 | 2026-10-17 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 77.5 | 911.7 | – | – | 32768 | 2026-10-11 | community | – | source |
| M4 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 77.5 | 911.7 | – | – | 32768 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 77.1 | 2436.0 | – | – | 65536 | 2026-10-09 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 77.0 | 2464.0 | – | – | 65536 | 2026-10-14 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 76.7 | 1622.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 76.7 | 1622.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 76.5 | 1745.0 | – | – | 4096 | 2026-10-09 | community | contradiction | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 76.3 | 1644.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | bf16 | oMLX | 76.3 | 1644.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 73.0 | 654.2 | – | – | 1024 | 2026-10-17 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 71.8 | 752.1 | – | – | 4096 | 2026-10-16 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 71.8 | 670.8 | – | – | 4096 | 2026-10-17 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 70.1 | 1101.0 | – | – | 65536 | 2026-10-10 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 69.6 | 655.9 | – | – | 32768 | 2026-10-14 | community | – | source |
| M1 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 68.6 | 722.4 | – | – | 8192 | 2026-10-16 | community | – | source |
| M4 Max | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 65.5 | 745.8 | – | – | 131072 | 2026-10-09 | community | – | source |
| M4 | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 64.2 | 606.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 61.8 | 479.6 | – | – | 1024 | 2026-10-16 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 61.8 | 479.6 | – | – | 1024 | 2026-10-16 | community | – | source |
| M5 Pro | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 61.7 | 1866.0 | – | – | 131072 | 2026-10-09 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 59.8 | 453.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 59.8 | 453.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M4 | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 59.0 | 557.6 | – | – | 4096 | 2026-10-16 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ4e | oMLX | 58.3 | 531.6 | – | – | 8192 | 2026-10-14 | community | – | source |
| M3 | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 57.3 | 554.7 | – | – | 1024 | 2026-10-12 | community | – | source |
| M2 Max | Mellum2.1-12B-A2.5B-Thinking | oQ8e | oMLX | 56.1 | 798.9 | – | – | 131072 | 2026-10-10 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 52.7 | 424.9 | – | – | 16384 | 2026-10-16 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 52.7 | 424.9 | – | – | 16384 | 2026-10-16 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ4 | oMLX | 42.8 | 345.9 | – | – | 4096 | 2026-10-17 | community | – | source |
| M1 Pro | Mellum2.1-12B-A2.5B-Thinking | oQ4 | oMLX | 37.9 | 313.4 | – | – | 16384 | 2026-10-17 | community | – | source |
| M4 | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 35.9 | 449.4 | – | – | 4096 | 2026-10-10 | community | – | source |
| M1 | Mellum2.1-12B-A2.5B-Thinking | mxfp4 | oMLX | 34.1 | 242.8 | – | – | 1024 | 2026-10-13 | community | – | source |
| M4 | Mellum2.1-12B-A2.5B-Thinking | 6bit | oMLX | 30.4 | 371.7 | – | – | 16384 | 2026-10-10 | community | – | source |