Qwen 3 VL 4B Instruct
Qwen 3 VL 4B Instruct is a chat model developed by Qwen, capable of processing up to 262,144 tokens and handling various inputs including text and images. It excels in multimodal reasoning, particularly in STEM and math-related tasks, and features advanced spatial perception and visual recognition capabilities.
Input
Output
Context
262K
Max Output
-
Parameters
4.4B
Technical Specifications
Model TypeChat
Context Window262,144 tokens
Max Output TokensNot available
Parameters4.4B
Release DateOct 11, 2025
Training CutoffSep 22, 2025
Licenseapache-2.0
Open SourceYes
Input Modalities
TextImage
Output Modalities
Text
Capabilities
Benchmarks
Artificial Analysis
3.7%
HLE
37.1%
GPQA
13.7%
SciCode
63.4%
MMLU Pro
37.0
Math
4.6
Coding
29.0%
LiveCodeBench
4.1
Intelligence
Resources & Links
Estimated GPU Requirements for Qwen 3 VL 4B Instruct
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Jun 24, 2026
Start building with the right model.
Automatically route workloads to the right model for every task, every time.