Alibaba: qwen3-vl-plus-2025-12-19 API
The definitive year-end snapshot of the Qwen3-VL series released by Alibaba Cloud on December 19, 2025. This version serves as a critical iteration for qwen3-vl-plus, optimized for top-tier multimodal reasoning and agentic execution. It fully integrates the DeepStack visual enhancement architecture and Interleaved-MRoPE, supporting interleaved text, image, and video inputs within a massive 256k context window. The 1219 snapshot specifically bolsters video timestamp alignment and 3D grounding, enabling precise localization of fleeting actions in videos or depth-aware object relations in static scenes.
- Context window: 262,144 tokens
- Max output: 32,768 tokens
- Input: text, image
- Output: text
- Reasoning: Supported
- Tool calling: Supported
- Released: 2025-12-19
- Knowledge cutoff: 2025-10-31
Frequently Asked Questions
What is the context window of qwen3-vl-plus-2025-12-19?
qwen3-vl-plus-2025-12-19 supports a context window of up to 262,144 tokens.
Does qwen3-vl-plus-2025-12-19 support function calling?
Yes. qwen3-vl-plus-2025-12-19 supports tool / function calling.
Does qwen3-vl-plus-2025-12-19 support reasoning?
Yes. qwen3-vl-plus-2025-12-19 is a reasoning-capable model.
What is the knowledge cutoff of qwen3-vl-plus-2025-12-19?
The knowledge cutoff of qwen3-vl-plus-2025-12-19 is 2025-10-31.