The AI arena is free today

Open Superagent

Qwen2.5 VL 72B Instruct vs Qwen3 VL 4B Instruct

Qwen2.5 VL 72B Instruct and Qwen3 VL 4B Instruct are closely matched at 12.8 and 10.8 on the LLM Stats Score.

Alibaba Cloud / Qwen Team · Alibaba Cloud / Qwen Team · Updated for 2026

Which is better?

Qwen2.5 VL 72B Instruct and Qwen3 VL 4B Instruct are closely matched on the overall LLM Stats Score at 12.8 and 10.8.

In the 15 individual benchmarks reported for both models, Qwen3 VL 4B Instruct wins 9; this is a narrower head-to-head signal than the composite indexes.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose Qwen2.5 VL 72B Instruct

  • your work emphasizes reasoning — it leads those capability indexes

Choose Qwen3 VL 4B Instruct

  • you value its reported benchmark strengths — it wins 9 of 15 exact shared results
  • you want the most recent training data — it shipped Sep 2025

At a glance

The differences that matter most.

Qwen2.5 VL 72B Instruct
Alibaba Cloud / Qwen Team
Qwen3 VL 4B Instruct
Alibaba Cloud / Qwen Team
Core performance indexes
LLM Stats Score
12.8
#238
10.8
#251
Reasoning
12.9
#229
8.4
#263
Agents
5.1
#137
2.8
#154
Cost, coverage & limits
Benchmark wins
6 of 15
9 of 15
Input price
— / M
$0.10 / M
Output price
— / M
$0.60 / M
Context window
—
262,144

Capability indexes

Additional strengths measured across groups of related public benchmarks

3 shared
Index
Qwen2.5 VL 72B Instruct
Qwen3 VL 4B Instruct
Math
7.9#255
8.2#253
Vision
9.8#121
7.7#137
Multimodal
12.7#102
9.3#119
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

30 reported for Qwen2.5 VL 72B Instruct · 45 for Qwen3 VL 4B Instruct

15 shared

Qwen2.5 VL 72B Instruct outperforms in 6 benchmarks (AI2D, CC-OCR, MathVista-Mini, MMStar, MVBench, OCRBench), while Qwen3 VL 4B Instruct is better at 9 benchmarks (Hallusion Bench, LVBench, MathVision, MLVU-M, MMMU-Pro, OCRBench-V2 (en), OSWorld, ScreenSpot, ScreenSpot Pro).

Qwen3 VL 4B Instruct has a slight edge in benchmark performance.

Sun Sep 06 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

68.0B diff

Qwen2.5 VL 72B Instruct has 68.0B more parameters than Qwen3 VL 4B Instruct, making it 1700.0% larger.

Alibaba Cloud / Qwen Team
Qwen2.5 VL 72B Instruct
72.0Bparameters
Alibaba Cloud / Qwen Team
Qwen3 VL 4B Instruct
4.0Bparameters
72.0B
Qwen2.5 VL 72B Instruct
4.0B
Qwen3 VL 4B Instruct

Context Window

Maximum input and output token capacity

Only Qwen3 VL 4B Instruct specifies input context (262,144 tokens). Only Qwen3 VL 4B Instruct specifies output context (262,144 tokens).

Alibaba Cloud / Qwen Team
Qwen2.5 VL 72B Instruct
Input- tokens
Output- tokens
Alibaba Cloud / Qwen Team
Qwen3 VL 4B Instruct
Input262,144 tokens
Output262,144 tokens
Sun Sep 06 2026 • llm-stats.com

Input capabilities

Documented input modalities across available providers

Both Qwen2.5 VL 72B Instruct and Qwen3 VL 4B Instruct support multimodal inputs.

They are both capable of processing various types of data, offering versatility in application.

Qwen2.5 VL 72B Instruct

Text
Images
Audio
Video

Qwen3 VL 4B Instruct

Text
Images
Audio
Video

License

Usage and distribution terms

Qwen2.5 VL 72B Instruct is licensed under tongyi-qianwen, while Qwen3 VL 4B Instruct uses Apache 2.0.

License differences may affect how you can use these models in commercial or open-source projects.

Qwen2.5 VL 72B Instruct

tongyi-qianwen

Open weights

Qwen3 VL 4B Instruct

Apache 2.0

Open weights

Release Timeline

When each model was launched

Qwen2.5 VL 72B Instruct was released on 2025-01-26, while Qwen3 VL 4B Instruct was released on 2025-09-22.

Qwen3 VL 4B Instruct is 8 months newer than Qwen2.5 VL 72B Instruct.

Qwen2.5 VL 72B Instruct

Jan 26, 2025

1.6 years ago

Qwen3 VL 4B Instruct

Sep 22, 2025

11 months ago

7mo newer

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Judge for yourself.

Run your own prompts against Qwen2.5 VL 72B Instruct and Qwen3 VL 4B Instruct side-by-side, then vote on the output you prefer.

Qwen2.5 VL 72B Instruct
✓ Preferred
Qwen3 VL 4B Instruct
Open in Playground

FAQ

Common questions about Qwen2.5 VL 72B Instruct vs Qwen3 VL 4B Instruct.

Which is better, Qwen2.5 VL 72B Instruct or Qwen3 VL 4B Instruct?

Qwen2.5 VL 72B Instruct and Qwen3 VL 4B Instruct are closely matched on the LLM Stats Score at 12.8 and 10.8. Qwen2.5 VL 72B Instruct is made by Alibaba Cloud / Qwen Team and Qwen3 VL 4B Instruct is made by Alibaba Cloud / Qwen Team. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does Qwen2.5 VL 72B Instruct compare to Qwen3 VL 4B Instruct in benchmarks?

Qwen2.5 VL 72B Instruct scores DocVQA: 96.4%, Android Control Low_EM: 93.7%, ChartQA: 89.5%, OCRBench: 88.5%, AI2D: 88.4%. Qwen3 VL 4B Instruct scores DocVQAtest: 95.3%, ScreenSpot: 94.0%, OCRBench: 88.1%, MMBench-V1.1: 85.1%, AI2D: 84.1%.

What are the context window sizes for Qwen2.5 VL 72B Instruct and Qwen3 VL 4B Instruct?

Qwen2.5 VL 72B Instruct supports an unknown number of tokens and Qwen3 VL 4B Instruct supports 262K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Qwen2.5 VL 72B Instruct and Qwen3 VL 4B Instruct?

Key differences include LLM Stats Score (12.8 vs 10.8), licensing (tongyi-qianwen vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.

Related comparisons