Qwen2.5-VL-7B-Instruct
10 records across 4 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M1 Max | Qwen2.5-VL-7B-Instruct | 4bit | oMLX | 65.7 | 449.4 | – | – | 1024 | 2026-10-07 | community | – | source |
| M1 Max | Qwen2.5-VL-7B-Instruct | 4bit | oMLX | 59.5 | 436.3 | – | – | 4096 | 2026-10-07 | community | – | source |
| M4 Max | Qwen2.5-VL-7B-Instruct | 8bit | oMLX | 57.4 | 939.1 | – | – | 4096 | 2026-10-17 | community | – | source |
| M4 Max | Qwen2.5-VL-7B-Instruct | 8bit | oMLX | 57.2 | 952.4 | – | – | 4096 | 2026-10-17 | community | – | source |
| M4 Max | Qwen2.5-VL-7B-Instruct | 8bit | oMLX | 51.8 | 753.0 | – | – | 16384 | 2026-10-17 | community | – | source |
| M4 Max | Qwen2.5-VL-7B-Instruct | 8bit | oMLX | 51.7 | 756.7 | – | – | 16384 | 2026-10-17 | community | – | source |
| M5 Pro | Qwen2.5-VL-7B-Instruct | 4bit | oMLX | 36.9 | 951.1 | – | – | 1024 | 2026-10-10 | community | – | source |
| M5 Pro | Qwen2.5-VL-7B-Instruct | 4bit | oMLX | 35.8 | 969.4 | – | – | 4096 | 2026-10-10 | community | – | source |
| M1 | Qwen2.5-VL-7B-Instruct | 4bit | oMLX | 13.7 | 145.0 | – | – | 1024 | 2026-10-07 | community | – | source |
| M1 | Qwen2.5-VL-7B-Instruct | 4bit | oMLX | 13.0 | 140.6 | – | – | 4096 | 2026-10-07 | community | – | source |