qwen2-5-vl-32b-instruct
Qwen builds the Qwen2.5 VL 32B Instruct model, a vision model capable of processing text and image inputs. It is genuinely best at handling complex tasks such as extended thinking and function calling, and can also produce structured output.
Input
Output
Context
128K
Max Output
16K
Parameters
32B
Technical Specifications
Model TypeVision
Context Window128,000 tokens
Max Output Tokens16,000 tokens
Parameters32B
Release DateMar 24, 2025
Training CutoffJun 1, 2024
Licenseapache-2.0
Open SourceYes
Input Modalities
TextImage
Output Modalities
Text
Capabilities
Resources & Links
Estimated GPU Requirements for qwen2-5-vl-32b-instruct
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
More Models by Qwen
qwen3-235b-a22b-thinking-2507262K context · 235B · Open weightsQwen 3 4B Thinking 2507262K context · 4B · Open weightsQwen 2.5 32B Instruct33K context · 32.8B · Open weightsQwen 2.5 14B Instruct AWQ33K context · 14.8B · Open weightsQwen 3 VL 8B Instruct262K context · 8.8B · Open weightsQwen 3.5 9B262K context · 9.7B · Open weights
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Jul 30, 2026
Your AI stack shouldn't stand still.
Every month new models become cheaper, faster, and more capable. Inferbase ensures your application automatically benefits without changing a single API call.