Qwen3.5-4B
436 records across 21 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 170.1 | 2063.0 | – | – | 1024 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 170.1 | 2063.0 | – | – | 1024 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 164.1 | 1977.0 | – | – | 1024 | 2026-10-09 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 163.6 | 2172.0 | – | – | 4096 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 163.6 | 1914.0 | – | – | 1024 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 163.6 | 2172.0 | – | – | 4096 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 163.6 | 1914.0 | – | – | 1024 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 159.3 | 2256.0 | – | – | 4096 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 159.3 | 2256.0 | – | – | 4096 | 2026-10-06 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 4bit | oMLX | 158.4 | 2159.0 | – | – | 4096 | 2026-10-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 143.9 | 1368.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 143.8 | 1358.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.5 | 1428.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.5 | 1428.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.5 | 1428.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.5 | 1428.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.5 | 1428.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.5 | 1428.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.5 | 1285.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 137.3 | 1278.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 132.4 | 1224.0 | – | – | 8192 | 2026-10-16 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 131.4 | 1220.0 | – | – | 8192 | 2026-10-16 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 128.2 | 1440.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 128.2 | 1440.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 128.2 | 1440.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 128.2 | 1440.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 128.2 | 1440.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 128.2 | 1440.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M1 Ultra | Qwen3.5-4B | oQ4e-fp16 | oMLX | 121.0 | 1889.0 | – | – | 8192 | 2026-10-10 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 117.4 | 1414.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 117.4 | 1414.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 117.4 | 1414.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 117.4 | 1414.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 117.4 | 1414.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 117.4 | 1414.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M4 Max | Qwen3.5-4B | 4bit | oMLX | 117.3 | 1315.0 | – | – | 1024 | 2026-10-12 | community | – | source |
| M4 Max | Qwen3.5-4B | 4bit | oMLX | 112.8 | 1066.0 | – | – | 4096 | 2026-10-12 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 8bit | oMLX | 111.5 | 1976.0 | – | – | 1024 | 2026-10-09 | community | – | source |
| M3 Ultra | Qwen3.5-4B | 8bit | oMLX | 109.5 | 2122.0 | – | – | 4096 | 2026-10-09 | community | – | source |
| M4 Max | Qwen3.5-4B | 4bit | oMLX | 108.5 | 1082.0 | – | – | 8192 | 2026-10-12 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 107.3 | 909.9 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 107.3 | 909.9 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 107.1 | 905.8 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 107.1 | 905.8 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 106.7 | 899.5 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 106.7 | 899.5 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 106.7 | 911.0 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 106.7 | 911.0 | – | – | 1024 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 102.2 | 736.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 102.2 | 736.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 102.2 | 736.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 102.2 | 736.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 102.2 | 736.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 102.2 | 736.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 99.9 | 1423.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 99.9 | 1423.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 99.9 | 1423.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 99.9 | 1423.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 99.9 | 1423.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 99.9 | 1423.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 99.9 | 842.7 | – | – | 4096 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 99.9 | 842.7 | – | – | 4096 | 2026-10-11 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 98.7 | 1355.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 98.7 | 1355.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 98.7 | 1355.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 98.7 | 1355.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 98.7 | 1355.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 98.7 | 1355.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 98.7 | 2579.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| M2 Max | Qwen3.5-4B | oQ4e | oMLX | 97.8 | 651.3 | – | – | 1024 | 2026-10-17 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 96.2 | 1448.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 96.2 | 1448.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 96.2 | 1448.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 96.2 | 1448.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 96.2 | 1448.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 96.2 | 1448.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 95.4 | 734.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 95.4 | 734.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 95.4 | 734.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 95.4 | 734.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 95.4 | 734.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 95.4 | 734.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 94.9 | 2386.0 | – | – | 1024 | 2026-10-15 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 94.9 | 2386.0 | – | – | 1024 | 2026-10-15 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 94.4 | 2351.0 | – | – | 1024 | 2026-10-15 | community | – | source |
| M2 Max | Qwen3.5-4B | oQ4e | oMLX | 94.0 | 682.6 | – | – | 4096 | 2026-10-17 | community | – | source |
| M2 Max | Qwen3.5-4B | oQ4e | oMLX | 93.4 | 718.1 | – | – | 1024 | 2026-10-15 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 92.6 | 3159.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 90.6 | 2313.0 | – | – | 4096 | 2026-10-15 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 90.1 | 1421.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 90.1 | 1421.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 90.1 | 1421.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 90.1 | 1421.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 90.1 | 1421.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 90.1 | 1421.0 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | oQ4e | oMLX | 90.1 | 636.9 | – | – | 8192 | 2026-10-17 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 90.0 | 2449.0 | – | – | 4096 | 2026-10-15 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 90.0 | 2449.0 | – | – | 4096 | 2026-10-15 | community | – | source |
| M2 Max | Qwen3.5-4B | oQ4e | oMLX | 89.5 | 691.2 | – | – | 4096 | 2026-10-15 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 89.3 | 842.5 | – | – | 4096 | 2026-10-11 | community | contradiction | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 89.3 | 842.5 | – | – | 4096 | 2026-10-11 | community | contradiction | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 87.0 | 720.3 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 87.0 | 720.3 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 87.0 | 720.3 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 87.0 | 720.3 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 87.0 | 720.3 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 87.0 | 720.3 | – | – | 8192 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 86.9 | 849.0 | – | – | 4096 | 2026-10-11 | community | contradiction | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 86.9 | 849.0 | – | – | 4096 | 2026-10-11 | community | contradiction | source |
| M2 Max | Qwen3.5-4B | oQ4e | oMLX | 84.8 | 643.0 | – | – | 8192 | 2026-10-15 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 84.0 | 2919.0 | – | – | 16384 | 2026-10-16 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 83.7 | 2488.0 | – | – | 8192 | 2026-10-15 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 83.7 | 2488.0 | – | – | 8192 | 2026-10-15 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 83.3 | 835.3 | – | – | 1024 | 2026-10-10 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 82.1 | 811.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 82.1 | 811.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 82.1 | 811.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 82.1 | 811.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 82.1 | 811.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 82.1 | 811.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 80.1 | 1362.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 80.1 | 1362.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 80.1 | 1362.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 80.1 | 1362.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 80.1 | 1362.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 80.1 | 1362.0 | – | – | 16384 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 80.1 | 843.0 | – | – | 4096 | 2026-10-10 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 79.8 | 836.1 | – | – | 4096 | 2026-10-11 | community | contradiction | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 79.8 | 836.1 | – | – | 4096 | 2026-10-11 | community | contradiction | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 78.1 | 666.7 | – | – | 1024 | 2026-10-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 76.8 | 798.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 76.8 | 798.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 76.8 | 798.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 76.8 | 798.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 76.8 | 798.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 76.8 | 798.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | oQ4e | oMLX | 76.7 | 575.3 | – | – | 16384 | 2026-10-15 | community | – | source |
| M4 Pro | Qwen3.5-4B | oQ4e | oMLX | 75.6 | 655.8 | – | – | 1024 | 2026-10-15 | community | – | source |
| M4 Pro | Qwen3.5-4B | oQ4e | oMLX | 75.0 | 687.7 | – | – | 4096 | 2026-10-15 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 74.5 | 1247.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 74.5 | 1247.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 74.5 | 1247.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 74.5 | 1247.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 74.5 | 1247.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 74.5 | 1247.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 73.5 | 2358.0 | – | – | 16384 | 2026-10-15 | community | – | source |
| M5 Pro | Qwen3.5-4B | 4bit | oMLX | 73.5 | 2358.0 | – | – | 16384 | 2026-10-15 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 73.4 | 691.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 73.4 | 691.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 73.4 | 691.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 73.4 | 691.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 73.4 | 691.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 73.4 | 691.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 72.4 | 781.9 | – | – | 8192 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 72.4 | 781.9 | – | – | 8192 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 72.4 | 781.9 | – | – | 8192 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 72.4 | 781.9 | – | – | 8192 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 72.4 | 781.9 | – | – | 8192 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 72.4 | 781.9 | – | – | 8192 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 72.4 | 653.1 | – | – | 4096 | 2026-10-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 71.6 | 589.6 | – | – | 4096 | 2026-10-11 | community | – | source |
| M4 Pro | Qwen3.5-4B | oQ4e | oMLX | 71.4 | 672.5 | – | – | 8192 | 2026-10-15 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 64.5 | 597.7 | – | – | 16384 | 2026-10-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | oQ4e | oMLX | 64.5 | 629.6 | – | – | 16384 | 2026-10-15 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 63.0 | 1255.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 63.0 | 1255.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 63.0 | 1255.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 63.0 | 1255.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 63.0 | 1255.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 63.0 | 1255.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 62.8 | 737.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 62.8 | 737.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 62.8 | 737.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 62.8 | 737.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 62.8 | 737.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 62.8 | 737.3 | – | – | 16384 | 2026-03-09 | community | – | source |
| M5 Pro | Qwen3.5-4B | 8bit | oMLX | 57.2 | 2105.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| RTX 2060 6 GB | Qwen3.5-4B | Q4_K_M | llama.cpp (CUDA) | 56.1 | – | – | – | 16384 | 2026-09-27 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 56.0 | 521.5 | – | – | 16384 | 2026-10-11 | community | contradiction | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 55.8 | 361.0 | – | – | 1024 | 2026-10-08 | community | – | source |
| M5 Pro | Qwen3.5-4B | 8bit | oMLX | 55.4 | 2643.0 | – | – | 4096 | 2026-10-16 | community | – | source |
| M3 Ultra | Qwen3.5-4B | unknown | oMLX | 54.3 | 1378.0 | – | – | 65536 | 2026-10-08 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 54.2 | 339.9 | – | – | 4096 | 2026-10-16 | community | – | source |
| M3 Max | Qwen3.5-4B | 4bit | oMLX | 54.0 | 624.8 | – | – | 1024 | 2026-10-16 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 53.8 | 292.6 | – | – | 1024 | 2026-03-09 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 53.8 | 292.6 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 53.7 | 534.2 | – | – | 32768 | 2026-10-09 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 53.2 | 341.5 | – | – | 4096 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | oQ4e | oMLX | 52.9 | 408.1 | – | – | 8192 | 2026-10-13 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 51.1 | 318.8 | – | – | 8192 | 2026-10-16 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 50.4 | 1051.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 50.4 | 1051.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 50.4 | 1051.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 50.4 | 1051.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 50.4 | 1051.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 4bit | oMLX | 50.4 | 1051.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | oQ4e | oMLX | 50.4 | 414.8 | – | – | 1024 | 2026-10-13 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 49.9 | 655.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 49.9 | 655.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 49.9 | 655.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 49.9 | 655.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 49.9 | 655.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 49.9 | 655.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 49.3 | 292.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 49.3 | 292.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 48.8 | 1261.0 | – | – | 16384 | 2026-10-08 | community | – | source |
| M3 Max | Qwen3.5-4B | 4bit | oMLX | 48.6 | 588.2 | – | – | 4096 | 2026-10-16 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 47.5 | 253.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 47.5 | 253.0 | – | – | 1024 | 2026-03-09 | community | – | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 47.4 | 1832.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 47.4 | 1832.0 | – | – | 4096 | 2026-10-11 | community | – | source |
| M1 Max | Qwen3.5-4B | 8bit | oMLX | 47.3 | 501.9 | – | – | 1024 | 2026-10-14 | community | – | source |
| RTX 2060 6 GB | Qwen3.5-4B | Q4_K_M | llama.cpp (Vulkan) | 47.0 | – | – | – | 16384 | 2026-09-27 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 46.7 | 1474.0 | – | – | 1024 | 2026-10-13 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 46.7 | 1463.0 | – | – | 1024 | 2026-10-13 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 46.5 | 255.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 46.5 | 255.0 | – | – | 4096 | 2026-03-09 | community | – | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 46.4 | 1759.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 46.4 | 1759.0 | – | – | 16384 | 2026-10-11 | community | – | source |
| M1 Max | Qwen3.5-4B | 8bit | oMLX | 46.1 | 490.5 | – | – | 4096 | 2026-10-14 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 45.5 | 1056.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 45.5 | 1056.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 45.5 | 1056.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 45.5 | 1056.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 45.5 | 1056.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Ultra | Qwen3.5-4B | 8bit | oMLX | 45.5 | 1056.0 | – | – | 65536 | 2026-03-09 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 45.5 | 282.3 | – | – | 16384 | 2026-10-16 | community | – | source |
| M4 | Qwen3.5-4B | oQ4e | oMLX | 45.4 | 419.9 | – | – | 4096 | 2026-10-13 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 44.9 | 1575.0 | – | – | 4096 | 2026-10-13 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 44.8 | 1366.0 | – | – | 4096 | 2026-10-07 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 44.8 | 1366.0 | – | – | 4096 | 2026-10-07 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 44.8 | 1577.0 | – | – | 4096 | 2026-10-13 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 44.7 | 1323.0 | – | – | 8192 | 2026-10-13 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 43.4 | 1253.0 | – | – | 4096 | 2026-10-10 | community | – | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 42.7 | 276.1 | – | – | 1024 | 2026-10-06 | community | contradiction | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 42.0 | 1345.0 | – | – | 8192 | 2026-10-08 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 41.8 | 1309.0 | – | – | 4096 | 2026-10-08 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 40.2 | 1204.0 | – | – | 16384 | 2026-10-13 | community | contradiction | source |
| M1 Pro | Qwen3.5-4B | 4bit | oMLX | 39.9 | 295.1 | – | – | 4096 | 2026-10-06 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 39.4 | 432.9 | – | – | 1024 | 2026-10-16 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 39.2 | 1057.0 | – | – | 1024 | 2026-10-08 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 39.0 | 429.5 | – | – | 1024 | 2026-10-14 | community | – | source |
| M1 Max | Qwen3.5-4B | 4bit | oMLX | 39.0 | 396.9 | – | – | 4096 | 2026-10-16 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 38.8 | 400.3 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 38.7 | 429.2 | – | – | 1024 | 2026-10-10 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 38.6 | 407.7 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 38.6 | 407.7 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 38.3 | 428.0 | – | – | 1024 | 2026-10-15 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.9 | 406.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.9 | 406.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.9 | 427.9 | – | – | 1024 | 2026-10-06 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.7 | 439.3 | – | – | 4096 | 2026-10-14 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.7 | 427.8 | – | – | 1024 | 2026-10-15 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.6 | 426.3 | – | – | 1024 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.6 | 438.1 | – | – | 4096 | 2026-10-16 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.4 | 406.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.4 | 400.4 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.4 | 406.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.4 | 406.1 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.4 | 429.0 | – | – | 1024 | 2026-10-08 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 37.3 | 1383.0 | – | – | 16384 | 2026-10-07 | community | contradiction | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 37.3 | 1383.0 | – | – | 16384 | 2026-10-07 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.2 | 423.8 | – | – | 4096 | 2026-10-10 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.1 | 397.8 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 37.1 | 397.8 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.9 | 400.6 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.8 | 567.7 | – | – | 4096 | 2026-10-15 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.6 | 405.2 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.6 | 405.2 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.6 | 568.7 | – | – | 4096 | 2026-10-15 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.5 | 435.8 | – | – | 4096 | 2026-10-06 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.4 | 396.6 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.4 | 396.6 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.4 | 561.4 | – | – | 4096 | 2026-10-07 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 36.3 | 534.5 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 36.3 | 534.5 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 36.3 | 534.5 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 36.3 | 534.5 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 36.3 | 534.5 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 Pro | Qwen3.5-4B | 4bit | oMLX | 36.3 | 534.5 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.3 | 432.9 | – | – | 4096 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.3 | 437.5 | – | – | 4096 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.3 | 415.0 | – | – | 1024 | 2026-10-13 | community | – | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 36.2 | 1091.0 | – | – | 16384 | 2026-10-10 | community | contradiction | source |
| M1 Max | Qwen3.5-4B | 8bit | oMLX | 36.2 | 354.2 | – | – | 32768 | 2026-10-14 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.1 | 427.3 | – | – | 8192 | 2026-10-16 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.0 | 403.8 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.0 | 403.8 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 36.0 | 403.8 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 35.5 | 403.4 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 35.5 | 403.4 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 35.1 | 378.3 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 35.1 | 378.3 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | oQ4e | oMLX | 34.9 | 436.7 | – | – | 4096 | 2026-10-14 | community | contradiction | source |
| M4 | Qwen3.5-4B | oQ4e-fp16 | oMLX | 34.8 | 437.0 | – | – | 4096 | 2026-10-06 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.7 | 427.3 | – | – | 4096 | 2026-10-07 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.7 | 420.1 | – | – | 8192 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | oQ4e-fp16 | oMLX | 34.7 | 437.5 | – | – | 4096 | 2026-10-11 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.5 | 391.0 | – | – | 1024 | 2026-10-06 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.5 | 391.0 | – | – | 1024 | 2026-10-06 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.5 | 424.2 | – | – | 8192 | 2026-10-06 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.5 | 425.6 | – | – | 8192 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.2 | 393.2 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.2 | 393.2 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.2 | 396.2 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.2 | 396.2 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.1 | 380.6 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.1 | 380.6 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 34.1 | 380.4 | – | – | 1024 | 2026-10-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 33.7 | 374.9 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 33.7 | 374.9 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 33.7 | 374.9 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 33.7 | 374.9 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 33.7 | 374.9 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 33.7 | 374.9 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.9 | 527.9 | – | – | 4096 | 2026-10-06 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.9 | 527.9 | – | – | 4096 | 2026-10-06 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.9 | 422.8 | – | – | 8192 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.8 | 403.6 | – | – | 4096 | 2026-10-06 | community | contradiction | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 32.7 | 1609.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 32.7 | 1609.0 | – | – | 32768 | 2026-10-11 | community | – | source |
| M3 | Qwen3.5-4B | 4bit | oMLX | 32.7 | 312.5 | – | – | 1024 | 2026-10-16 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.6 | 522.3 | – | – | 4096 | 2026-10-06 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.6 | 402.3 | – | – | 4096 | 2026-10-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.0 | 373.3 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.0 | 373.3 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.0 | 373.3 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.0 | 373.3 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.0 | 373.3 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 32.0 | 373.3 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 31.9 | 338.1 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 31.9 | 338.1 | – | – | 1024 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 31.8 | 385.2 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 31.8 | 385.2 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 31.8 | 385.2 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M3 | Qwen3.5-4B | 4bit | oMLX | 31.5 | 323.8 | – | – | 4096 | 2026-10-16 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 31.3 | 413.9 | – | – | 4096 | 2026-10-10 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 30.9 | 395.6 | – | – | 16384 | 2026-10-06 | community | – | source |
| M3 | Qwen3.5-4B | 4bit | oMLX | 30.8 | 334.5 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 30.7 | 380.9 | – | – | 16384 | 2026-10-08 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.8 | 327.4 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.8 | 327.4 | – | – | 4096 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.5 | 364.6 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.5 | 364.6 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M3 | Qwen3.5-4B | 4bit | oMLX | 29.5 | 296.7 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.5 | 364.6 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.5 | 364.6 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.5 | 364.6 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 29.5 | 364.6 | – | – | 8192 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | oQ4e-fp16 | oMLX | 29.3 | 428.7 | – | – | 8192 | 2026-10-11 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 28.6 | 576.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 28.6 | 576.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 28.6 | 576.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 28.6 | 576.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 28.6 | 576.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 28.6 | 576.0 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 27.6 | 371.1 | – | – | 16384 | 2026-10-10 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 27.1 | 463.8 | – | – | 16384 | 2026-10-06 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 25.2 | 340.6 | – | – | 16384 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 25.2 | 340.6 | – | – | 16384 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 25.2 | 340.6 | – | – | 16384 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 25.2 | 340.6 | – | – | 16384 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 25.2 | 340.6 | – | – | 16384 | 2026-03-09 | community | contradiction | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 25.2 | 340.6 | – | – | 16384 | 2026-03-09 | community | contradiction | source |
| M5 | Qwen3.5-4B | 4bit | oMLX | 23.6 | 912.8 | – | – | 32768 | 2026-10-13 | community | – | source |
| M4 | Qwen3.5-4B | OptiQ | oMLX | 23.3 | 254.9 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | OptiQ | oMLX | 23.3 | 254.9 | – | – | 1024 | 2026-03-09 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 22.8 | 181.2 | – | – | 1024 | 2026-10-10 | community | – | source |
| AMD Custom GPU 0932 | Qwen3.5-4B 4.21B | Q4_K - Medium | llama.cpp (Vulkan) | 22.67 | 190.20 | – | – | tg128 | 2026-08-20 | community | – | source |
| AMD Custom GPU 0932 | Qwen3.5-4B 4.21B | Q4_K - Medium | llama.cpp (Vulkan) | 22.55 | 189.22 | – | – | tg128 | 2026-08-20 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 22.2 | 145.3 | – | – | 1024 | 2026-03-09 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 22.2 | 145.3 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | OptiQ | oMLX | 22.1 | 263.4 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | OptiQ | oMLX | 22.1 | 263.4 | – | – | 4096 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 22.1 | 395.5 | – | – | 32768 | 2026-10-06 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 22.0 | 168.7 | – | – | 4096 | 2026-10-10 | community | – | source |
| M4 | Qwen3.5-4B | 8bit | oMLX | 21.8 | 425.5 | – | – | 1024 | 2026-10-11 | community | – | source |
| M4 | Qwen3.5-4B | 8bit | oMLX | 21.2 | 428.1 | – | – | 4096 | 2026-10-11 | community | – | source |
| M4 | Qwen3.5-4B | 8bit | oMLX | 21.1 | 421.7 | – | – | 4096 | 2026-10-08 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 21.0 | 144.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 21.0 | 144.5 | – | – | 4096 | 2026-03-09 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 20.6 | 149.5 | – | – | 4096 | 2026-10-13 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 20.3 | 299.7 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 20.3 | 299.7 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 20.3 | 299.7 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 20.3 | 299.7 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 20.3 | 299.7 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 20.3 | 299.7 | – | – | 32768 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 19.9 | 158.6 | – | – | 1024 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 8bit | oMLX | 19.2 | 382.1 | – | – | 16384 | 2026-10-08 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 18.5 | 166.3 | – | – | 4096 | 2026-10-14 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 18.2 | 468.7 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 18.2 | 468.7 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 18.2 | 468.7 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 18.2 | 468.7 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 18.2 | 468.7 | – | – | 65536 | 2026-03-09 | community | – | source |
| M2 Max | Qwen3.5-4B | 4bit | oMLX | 18.2 | 468.7 | – | – | 65536 | 2026-03-09 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 18.1 | 136.0 | – | – | 16384 | 2026-10-13 | community | – | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 17.5 | 774.8 | – | – | 131072 | 2026-10-11 | community | – | source |
| M6 | Qwen3.5-4B | 4bit | oMLX | 17.5 | 774.8 | – | – | 131072 | 2026-10-11 | community | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | 16.83 | – | – | – | 32768 | 2026-10-17 | sourced | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | 16.6 | – | – | – | 32768 | 2026-10-17 | sourced | – | source |
| M4 | Qwen3.5-4B | 8bit | oMLX | 16.5 | 330.3 | – | – | 32768 | 2026-10-08 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 15.6 | 130.2 | – | – | 16384 | 2026-10-14 | community | – | source |
| RTX 2060 6 GB | Qwen3.5-4B | Q4_K_M | llama.cpp (Vulkan) | 15.4 | 6.4 | – | – | 16384 | 2026-09-27 | community | – | source |
| OnePlus 13T | Qwen3.5-4B | Q4_0 | llama.cpp (Hexagon) | 14.47 | – | – | – | tg32 | 2026-09-28 | community | – | source |
| OnePlus 13T | Qwen3.5-4B | Q4_0 | llama.cpp (Hexagon) | 14.38 | – | – | – | tg32 | 2026-09-28 | community | – | source |
| M1 | Qwen3.5-4B | oQ8e | oMLX | 14.2 | 219.4 | – | – | 1024 | 2026-10-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 13.6 | 231.3 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 13.6 | 231.3 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 13.6 | 231.3 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 13.6 | 231.3 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 13.6 | 231.3 | – | – | 65536 | 2026-03-09 | community | – | source |
| M4 | Qwen3.5-4B | 4bit | oMLX | 13.6 | 231.3 | – | – | 65536 | 2026-03-09 | community | – | source |
| M1 | Qwen3.5-4B | oQ8e | oMLX | 13.6 | 231.1 | – | – | 4096 | 2026-10-09 | community | – | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 13.2 | 93.8 | – | – | 1024 | 2026-10-07 | community | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | 13.19 | – | – | – | 32768 | 2026-10-17 | sourced | contradiction | source |
| M1 | Qwen3.5-4B | 4bit | oMLX | 13.0 | 91.9 | – | – | 1024 | 2026-10-07 | community | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | 12.36 | – | – | – | 32768 | 2026-10-17 | sourced | contradiction | source |
| AMX | Qwen3.5-4B | BF16 | vLLM | 9.48 | – | – | – | 100000 | 2026-09-28 | community | – | source |
| AMX | Qwen3.5-4B | BF16 | vLLM | 9.36 | – | – | – | 100000 | 2026-09-28 | community | – | source |
| Snapdragon 7 Gen 4 (SM7750) | Qwen3.5-4B | Q4_0 | llama.cpp (CPU) | 9.1 | 46.3 | – | – | tg64 | 2026-09-26 | community | – | source |
| AMX | Qwen3.5-4B | BF16 | vLLM | 8.62 | – | – | – | 100000 | 2026-09-28 | community | – | source |
| Snapdragon 7 Gen 4 (SM7750) | Qwen3.5-4B | Q4_0 | llama.cpp (Hexagon) | 8.1 | 437.4 | – | – | tg64 | 2026-09-26 | community | – | source |
| AMX | Qwen3.5-4B | BF16 | vLLM | 8.01 | – | – | – | 100000 | 2026-09-28 | community | – | source |
| Snapdragon 7 Gen 4 (SM7750) | Qwen3.5-4B | Q4_0 | llama.cpp (OpenCL) | 7.5 | 77.5 | – | – | tg64 | 2026-09-26 | community | – | source |
| AMX | Qwen3.5-4B | BF16 | vLLM | 7.08 | – | – | – | 100000 | 2026-09-28 | community | – | source |
| AMX | Qwen3.5-4B | BF16 | vLLM | 6.73 | – | – | – | 100000 | 2026-09-28 | community | – | source |
| Snapdragon 7 Gen 4 (SM7750) | Qwen3.5-4B | Q4_0 | llama.cpp (Hexagon) | 6.2 | 25.6 | – | – | tg64 | 2026-09-26 | community | – | source |
| RTX 2060 6 GB | Qwen3.5-4B | Q4_K_M | llama.cpp (CUDA) | 3.09 | 0.44 | – | – | 16384 | 2026-09-27 | community | – | source |
| OnePlus 13T | Qwen3.5-4B | Q4_0 | llama.cpp (Hexagon) | – | 665.20 | – | – | – | 2026-09-28 | community | – | source |
| OnePlus 13T | Qwen3.5-4B | Q4_0 | llama.cpp (Hexagon) | – | 669.88 | – | – | – | 2026-09-28 | community | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | – | 82.46 | – | – | 32768 | 2026-10-17 | sourced | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | – | 116.44 | – | – | 32768 | 2026-10-17 | sourced | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | – | 73.78 | – | – | 32768 | 2026-10-17 | sourced | – | source |
| AMD Radeon HD 7900 | Qwen3.5-4B | Q4_K_M | llama.cpp | – | 116.24 | – | – | 32768 | 2026-10-17 | sourced | – | source |