What are the hardware requirements for Qwen2 VL 72B?

Qwen2 VL 72B requires approximately 146.8 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen2 VL 72B and how large is it?

Qwen2 VL 72B is developed by Qwen, a model with 72b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Qwen2 VL 72B?

Qwen2 VL 72B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Qwen2 VL 72B by Qwen — VRAM 146.8GB, 32K context

Name: Qwen2 VL 72B
Author: Qwen

Qwen2 VL 72B is an open-source language model by Qwen. Features: 72b LLM, VRAM: 146.8GB, Context: 32K, License: other.

Arxiv:2308.12966 Arxiv:2409.12191 Conversational Deploy:azure En Endpoints compatible Image-text-to-text Multimodal Qwen2 vl Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Qwen/Qwen2-VL-72B

Qwen2 VL 72B Parameters and Internals

LLM Name	Qwen2 VL 72B
Repository 🤗	https://huggingface.co/Qwen/Qwen2-VL-72B
Model Size	72b
Required VRAM	146.8 GB
Updated	2026-08-10
Maintainer	Qwen
Model Type	qwen2_vl
Model Files	3.9 GB: 1-of-38 3.8 GB: 2-of-38 4.0 GB: 3-of-38 4.0 GB: 4-of-38 4.0 GB: 5-of-38 3.8 GB: 6-of-38 4.0 GB: 7-of-38 4.0 GB: 8-of-38 4.0 GB: 9-of-38 3.8 GB: 10-of-38 4.0 GB: 11-of-38 4.0 GB: 12-of-38 4.0 GB: 13-of-38 3.8 GB: 14-of-38 4.0 GB: 15-of-38 4.0 GB: 16-of-38 4.0 GB: 17-of-38 3.8 GB: 18-of-38 4.0 GB: 19-of-38 4.0 GB: 20-of-38 4.0 GB: 21-of-38 3.8 GB: 22-of-38 4.0 GB: 23-of-38 4.0 GB: 24-of-38 4.0 GB: 25-of-38 3.8 GB: 26-of-38 4.0 GB: 27-of-38 4.0 GB: 28-of-38 4.0 GB: 29-of-38 3.8 GB: 30-of-38 4.0 GB: 31-of-38 4.0 GB: 32-of-38 4.0 GB: 33-of-38 3.8 GB: 34-of-38 4.0 GB: 35-of-38 4.0 GB: 36-of-38 2.2 GB: 37-of-38 2.5 GB: 38-of-38
Supported Languages	en
Model Architecture	Qwen2VLForConditionalGeneration
License	other
Context Length	32768
Model Max Length	32768
Transformers Version	4.41.2
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	152064
Torch Data Type	bfloat16
Errors	replace

Rank the Qwen2 VL 72B capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Quantized Models of the Qwen2 VL 72B

Model	Likes	Downloads	VRAM
QVQ 72B Preview 4bit	7	20	41 GB
QVQ 72B Preview 6bit	2	34	59 GB
QVQ 72B Preview 3bit	5	21	32 GB
QVQ 72B Preview 8bit	3	22	77 GB

Best Alternatives to Qwen2 VL 72B

Best Alternatives	Context / RAM	Downloads	Likes
QVQ 72B Preview Bf16	125K / 147.7 GB	23	3
QVQ 72B Preview	125K / 146.8 GB	8572	610
UI TARS 72B DPO	32K / 147.5 GB	1735	158
UI TARS 72B DPO Bf16	32K / 147.7 GB	30	1
UI TARS 72B SFT Bf16	32K / 147.7 GB	32	1
UI TARS 72B SFT	32K / 204.2 GB	89	25
Qwen2 VL 72B Instruct	32K / 146.8 GB	60157	311
QVQ 72B Preview 4bit	125K / 41.3 GB	20	7
QVQ 72B Preview 6bit	125K / 59.7 GB	34	2
QVQ 72B Preview 3bit	125K / 32.2 GB	21	5

Note: green Score (e.g. "73.2") means that the model is better than Qwen/Qwen2-VL-72B.