Model Comparison

Qwen2.5 VL 72B Instruct vs Qwen3 VL 8B Thinking

Qwen3 VL 8B Thinking shows notably better performance in the majority of benchmarks.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

15 benchmarks

Qwen2.5 VL 72B Instruct outperforms in 4 benchmarks (AI2D, CC-OCR, MVBench, OCRBench), while Qwen3 VL 8B Thinking is better at 11 benchmarks (Hallusion Bench, LVBench, MathVision, MathVista-Mini, MLVU-M, MMMU-Pro, MMStar, OCRBench-V2 (en), OSWorld, ScreenSpot, ScreenSpot Pro).

Qwen3 VL 8B Thinking shows notably better performance in the majority of benchmarks.

Wed May 13 2026 • llm-stats.com

Arena Performance

Human preference votes

Model Size

Parameter count comparison

63.0B diff

Qwen2.5 VL 72B Instruct has 63.0B more parameters than Qwen3 VL 8B Thinking, making it 700.0% larger.

Qwen2.5 VL 72B Instruct

72.0Bparameters

Qwen3 VL 8B Thinking

9.0Bparameters

72.0B

Qwen2.5 VL 72B Instruct

9.0B

Qwen3 VL 8B Thinking

Context Window

Maximum input and output token capacity

Only Qwen3 VL 8B Thinking specifies input context (262,144 tokens). Only Qwen3 VL 8B Thinking specifies output context (262,144 tokens).

Qwen2.5 VL 72B Instruct

Input- tokens

Output- tokens

Qwen3 VL 8B Thinking

Input262,144 tokens

Output262,144 tokens

Wed May 13 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Both Qwen2.5 VL 72B Instruct and Qwen3 VL 8B Thinking 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 8B Thinking

Text

Images

Audio

Video

License

Usage and distribution terms

Qwen2.5 VL 72B Instruct is licensed under tongyi-qianwen, while Qwen3 VL 8B Thinking 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 8B Thinking

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 8B Thinking was released on 2025-09-22.

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

Qwen2.5 VL 72B Instruct

Jan 26, 2025

1.3 years ago

Qwen3 VL 8B Thinking

Sep 22, 2025

7 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→

Key Takeaways

Qwen2.5 VL 72B Instruct

View details

Alibaba Cloud / Qwen Team

Higher AI2D score (88.4% vs 84.9%)

Higher CC-OCR score (79.8% vs 76.3%)

Higher MVBench score (70.4% vs 69.0%)

Higher OCRBench score (88.5% vs 81.9%)

Qwen3 VL 8B Thinking

View details

Alibaba Cloud / Qwen Team

Larger context window (262,144 tokens)

Higher Hallusion Bench score (65.4% vs 55.2%)

Higher LVBench score (55.8% vs 47.3%)

Higher MathVision score (62.7% vs 38.1%)

Higher MathVista-Mini score (81.4% vs 74.8%)

Higher MLVU-M score (75.1% vs 74.6%)

Higher MMMU-Pro score (60.4% vs 51.1%)

Higher MMStar score (75.3% vs 70.8%)

Higher OCRBench-V2 (en) score (63.9% vs 61.5%)

Higher OSWorld score (33.9% vs 8.8%)

Higher ScreenSpot score (93.6% vs 87.1%)

Higher ScreenSpot Pro score (46.6% vs 43.6%)

Detailed Comparison

AI Model Comparison Table
Feature	Qwen2.5 VL 72B Instruct	Qwen3 VL 8B Thinking

FAQ

Common questions about Qwen2.5 VL 72B Instruct vs Qwen3 VL 8B Thinking.

Which is better, Qwen2.5 VL 72B Instruct or Qwen3 VL 8B Thinking?

Qwen3 VL 8B Thinking shows notably better performance in the majority of benchmarks. Qwen2.5 VL 72B Instruct is made by Alibaba Cloud / Qwen Team and Qwen3 VL 8B Thinking is made by Alibaba Cloud / Qwen Team. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does Qwen2.5 VL 72B Instruct compare to Qwen3 VL 8B Thinking 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 8B Thinking scores DocVQAtest: 95.3%, ScreenSpot: 93.6%, MMLU-Redux: 88.8%, MMBench-V1.1: 87.5%, InfoVQAtest: 86.0%.

What are the context window sizes for Qwen2.5 VL 72B Instruct and Qwen3 VL 8B Thinking?

Qwen2.5 VL 72B Instruct supports an unknown number of tokens and Qwen3 VL 8B Thinking 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 8B Thinking?

Key differences include licensing (tongyi-qianwen vs Apache 2.0). See the full comparison above for benchmark-by-benchmark results.