Qwen 2.5 VL 72B Instruct — VRAM, speed & local setup
Written by Jakub Rusinowski · Last updated
The most capable open vision-language model. Beats GPT-4V on multiple visual benchmarks. Ideal for enterprise document processing, agentic vision tasks, and complex image reasoning.
Qwen 2.5 VL 72B Instruct needs about 45 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
VRAM and speed by quantization
Quoted against NVIDIA RTX 4090 (24 GB). Includes the KV cache at 8K context, so it reads higher than the headline figure.
| Quant | Memory | VRAM | Speed (est.) | Fit |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 27.6 GB | ~4 tok/s | Offload | |
| Q3_K_M 3.41 bpw | 34.8 GB | ~3 tok/s | Offload | |
| Q4_K_M 4.83 bpw | 47.8 GB | ~2 tok/s | Offload | |
| Q5_K_M 5.67 bpw | 55.5 GB | ~2 tok/s | Offload | |
| Q6_K 6.56 bpw | 63.7 GB | — | Too big | |
| Q8_0 8.50 bpw | 81.5 GB | — | Too big | |
| F16 16.00 bpw | 150.3 GB | — | Too big |
Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Qwen 2.5 VL 72B Instruct VRAM calculator →
Get Qwen 2.5 VL 72B Instruct running
The cheapest catalogued GPU that runs Qwen 2.5 VL 72B Instruct is the NVIDIA DGX Spark (128 GB).
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
How to run Qwen 2.5 VL 72B Instruct
Install Ollama, then run:
ollama run qwen2.5vl:72bSpecifications
Curated — A hand-written entry from before this catalogue recorded its sources. The figures are long-standing but their provenance is not on file.
- Parameters
- 72 Billion
- Context window
- 128,000
- Architecture
- Dense + ViT
- Provider
- Alibaba Cloud
- Licence
- Apache 2.0
- Practical deployment
- Workstation (large unified memory or pro GPU)
- Specified at
- Q4_K_M
- System RAM
- 80 GB
- Record updated
- 2025-01-25
Commercial use permitted. No usage restrictions beyond attribution.
Quality and use cases
Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.
Can I run Qwen 2.5 VL 72B Instruct on my GPU?
- Qwen 2.5 VL 72B Instruct on AMD Radeon RX 9060 XT 8GB
- Qwen 2.5 VL 72B Instruct on AMD Ryzen AI Max+ 395
- Qwen 2.5 VL 72B Instruct on Apple M1 Max
- Qwen 2.5 VL 72B Instruct on Apple M2 Max
- Qwen 2.5 VL 72B Instruct on Apple M3 Max
- Qwen 2.5 VL 72B Instruct on Apple M5 Pro
- Qwen 2.5 VL 72B Instruct on Intel Arc B570
- Qwen 2.5 VL 72B Instruct on Intel Arc B580
- Qwen 2.5 VL 72B Instruct on NVIDIA A100 80GB (PCIe)
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 3060 (12GB)
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 3070
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 3070 Ti
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 3080 (10GB)
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 4060
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 4070
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 4070 Super
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 4070 Ti
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 5060
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 5060 Ti 8GB
- Qwen 2.5 VL 72B Instruct on NVIDIA GeForce RTX 5070
- Qwen 2.5 VL 72B Instruct on NVIDIA H100 80GB (PCIe)
- Qwen 2.5 VL 72B Instruct on NVIDIA L40S
- Qwen 2.5 VL 72B Instruct on NVIDIA RTX 6000 Ada Generation
- Qwen 2.5 VL 72B Instruct on NVIDIA RTX PRO 6000 Blackwell
Other Qwen 2.5 VL sizes
Qwen 2.5 VL 72B Instruct — frequently asked questions
How much VRAM does Qwen 2.5 VL 72B Instruct need?
About 45 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.
Does Qwen 2.5 VL 72B Instruct run on an RTX 4090 (24 GB)?
No. Qwen 2.5 VL 72B Instruct needs about 45 GB at Q4_K_M, more than a single RTX 4090's 24 GB. It needs a larger card, several GPUs, or Apple Silicon with enough unified memory — or it runs with part of the weights offloaded to system RAM, which is much slower.
How do I run Qwen 2.5 VL 72B Instruct locally?
Install Ollama and run `ollama run qwen2.5vl:72b`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does Qwen 2.5 VL come in?
Qwen 2.5 VL 7B Instruct (6 GB), Qwen 2.5 VL 72B Instruct (45 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.