Qwen

Qwen2.5-VL 72B Instruct

CTX131K

23Quality

Vision · Tools

Input$2.80/M

Output$8.40/M

Qwen vision-language model for visual reasoning, documents, and agent tasks

Open weightsQwen license385/1044

Drag to turn

QwenTextRare

Qwen2.5-VL 72B Instruct

Qwen vision-language model for visual reasoning, documents, and agent tasks

qwen2-5-vl-72b-instruct

Type: Text
Quality: 23 · Top 77% on the LMArena Vision
Context window: 131,072 tokens
Max output: 8,192 tokens
Input price: $2.80 / 1M tokens
Output price: $8.40 / 1M tokens
Cache read: —
Cache write: —
Parameters: 73B (73,410,777,344)
Downloads: 67K in the last 30 days
Likes: 660
Access to weights: Open download
Released: Sep 1, 2024
Knowledge cutoff: Apr 2024
Takes in: Text, Image
Puts out: Text
Reasoning: No
Tool calling: Yes
Structured output: No
File attachments: No
Access: Open weights
License: Qwen license
Rarity: Rare
Set number: 385/1044

Open Qwen docs Weights on Hugging Face Quality 23 Where to run it 10 Evolution 2/3 Race it 32 tok/s Meet its creature Battle MiMo-V2.6-Pro-UltraSpeed

License read from the model’s Hugging Face card. Specs from models.dev, downloads from Hugging Face, last changed Oct 19, 2026. Official source: www.alibabacloud.com.

How good is it?

Public leaderboard scores for Qwen2.5-VL 72B Instruct, refreshed daily from Epoch AI and LMArena.

Top 77% of the 165 models on the LMArena Vision: #128.

Quality is the share of a leaderboard’s models this one beats, from 0 to 100. Chat models are rated on Epoch AI’s Capabilities Index when it lists them, otherwise on LMArena; image and video models on LMArena’s image and video arenas.

Leaderboards

LMArena Vision#128 of 165
Questions about images, voted by people.1,107 rating

Scores from LMArena, licensed CC BY 4.0. Boards list models under their own names, so a close match is shown with the name it was scored under. Credits

Where to run it

8 providers sell Qwen2.5-VL 72B Instruct through an API, and 2 Hugging Face hosts serve it. Prices are per million tokens, cheapest first.

CheapestNanoGPT$0.70 in · $0.70 out
From the labAlibaba$2.80 in · $8.40 out
Fastest on Hugging FaceOVHcloud32 tokens/s

API providers

Provider	Input	Output	Context
NanoGPTCheapest paid`TEE/qwen2.5-vl-72b-instruct`	$0.70	$0.70	66K
Kilo Gateway`qwen/qwen2.5-vl-72b-instruct`	$0.80	$1.00	128K
OpenRouter`qwen/qwen2.5-vl-72b-instruct`	$0.80	$1.00	128K
OVHcloud AI Endpoints`qwen2.5-vl-72b-instruct`	$1.01	$1.01	33K
Cortecs`qwen2.5-vl-72b-instruct`	$1.01	$1.01	32K
Alibaba (China)`qwen2-5-vl-72b-instruct`	$2.29	$6.88	131K
AlibabaOfficial`qwen2-5-vl-72b-instruct`	$2.80	$8.40	131K
Qiniu`qwen2.5-vl-72b-instruct`	—	—	128K

Hugging Face Inference Providers

Host	Input	Output	Context	Speed	First token
OVHcloudFastest	$1.01	$1.01	33K	32 tok/s	0.3s
Featherless	—	—	—	—	—

From models.dev and Hugging Face, checked daily. Resellers’ listings can lag behind their price pages, so confirm with the provider before you commit.

Face off

Current text models from other labs at a similar output price.