Qwen3 VL Plus

alibaba

The Qwen3 series of visual understanding models effectively integrates thinking and non-thinking modes. Compared to the snapshot released on September 23, this version delivers superior performance in reasoning and analysis tasks as well as style control, while also offering lower latency and faster response speeds. This version is based on a snapshot taken on December 19, 2025.

Try Now

Capabilities

Tool Use

Image Input

Example Use Cases

Balanced visual analysis

Visual reasoning at moderate cost

Image or video understanding with thinking

Technical Specifications

Context Window

262,144 tokens

Max Output

32,768 tokens

Cache Miss Cost

$0.20 per 1M tokens

Non-Reasoning Cost

$1.60 per 1M tokens

Web Search Cost

$15 per 1K calls

Code Execution Cost

$0.19 per 1K calls

⚠️ Legacy

Made legacy on

Reason

Untested

Recommended Replacement

Qwen3 Max