Qwen2.5-VL-3B-Instruct
4 records across 2 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M5 | Qwen2.5-VL-3B-Instruct | 4bit | oMLX | 63.4 | 1795.0 | – | – | 1024 | 2026-10-13 | community | – | source |
| M5 | Qwen2.5-VL-3B-Instruct | 4bit | oMLX | 57.4 | 1701.0 | – | – | 4096 | 2026-10-13 | community | – | source |
| M1 Pro | Qwen2.5-VL-3B-Instruct | 4bit | oMLX | 44.8 | 337.7 | – | – | 16384 | 2026-10-14 | community | – | source |
| M1 Pro | Qwen2.5-VL-3B-Instruct | 4bit | oMLX | 35.0 | 327.5 | – | – | 4096 | 2026-10-14 | community | – | source |