Skip to content

Qwen 2.5 VL 72B

by Alibaba · qwen-2.5 family

72B

parameters

text-generation vision reasoning multilingual math

Qwen 2.5 VL 72B is Alibaba's flagship vision-language model, offering top-tier multimodal performance at 72 billion parameters. It excels at complex visual reasoning, mathematical problem solving from images, document analysis, and multilingual visual understanding. This model delivers frontier-level vision capabilities with strong performance on benchmarks like MathVista and DocVQA, making it a powerful choice for demanding multimodal workloads when sufficient hardware is available.

Quick Start with Ollama

ollama run 72b-q4_K_M
Resources Ollama Hugging Face Official Page
Creator Alibaba
Parameters 72B
Architecture transformer-decoder
Context 32K tokens
Released Jan 26, 2025
License Apache 2.0
Ollama qwen2.5vl:72b

Quantization Options

Format File Size VRAM Required Quality Ollama Tag
Q4_K_M rec 36 GB 41 GB 72b-q4_K_M
Q8_0 72 GB 78 GB 72b-q8_0

Compatible Hardware

Q4_K_M requires 41 GB VRAM

Compatible Hardware

HardwareVRAMTypeFitEst. Speed
Mac Studio M4 Ultra 512GB512 GBmacRuns~20 tok/s
Mac Pro M2 Ultra 192GB192 GBmacRuns~20 tok/s
Mac Studio M4 Ultra 192GB192 GBmacRuns~20 tok/s
Mac Studio M4 Max 128GB128 GBmacRuns~13 tok/s
MacBook Pro M4 Max 128GB128 GBmacRuns~13 tok/s
MacBook Pro M5 Max 128GB128 GBmacRuns~13 tok/s
NVIDIA RTX PRO 6000 Blackwell96 GBgpuRuns~47 tok/s
MacBook Pro M3 Max 96GB96 GBmacRuns~10 tok/s
Mac mini M4 Pro 64GB64 GBmacRuns~7 tok/s
Mac Studio M4 Max 64GB64 GBmacRuns~13 tok/s
MacBook Pro M4 Max 64GB64 GBmacRuns~13 tok/s
MacBook Pro M5 Max 64GB64 GBmacRuns~13 tok/s
NVIDIA RTX 6000 Ada Generation48 GBgpuRuns (tight)~23 tok/s
NVIDIA RTX A600048 GBgpuRuns (tight)~19 tok/s
NVIDIA RTX PRO 5000 Blackwell48 GBgpuRuns (tight)~23 tok/s
Mac mini M4 Pro 48GB48 GBmacRuns (tight)~7 tok/s
MacBook Pro M3 Max 48GB48 GBmacRuns (tight)~10 tok/s
MacBook Pro M4 Pro 48GB48 GBmacRuns (tight)~7 tok/s
MacBook Pro M4 Max 48GB48 GBmacRuns (tight)~13 tok/s
MacBook Pro M5 Max 48GB48 GBmacRuns (tight)~10 tok/s
MacBook Pro M5 Pro 48GB48 GBmacRuns (tight)~7 tok/s
Mac Studio M4 Max 36GB36 GBmacCPU Offload~4 tok/s
MacBook Pro M3 Pro 36GB36 GBmacCPU Offload~1 tok/s
MacBook Pro M5 Max 36GB36 GBmacCPU Offload~3 tok/s
NVIDIA RTX 5000 Ada Generation32 GBgpuCPU Offload~5 tok/s
NVIDIA GeForce RTX 509032 GBgpuCPU Offload~13 tok/s
iMac M4 32GB32 GBmacCPU Offload~1 tok/s
Mac mini M4 32GB32 GBmacCPU Offload~1 tok/s
MacBook Air M4 32GB32 GBmacCPU Offload~1 tok/s
MacBook Air M5 32GB32 GBmacCPU Offload~1 tok/s
MacBook Pro M5 32GB32 GBmacCPU Offload~1 tok/s
76 hardware device(s) cannot run this model at Q4_K_M.

Benchmark Scores

85.0
mmlu