Jina-VLM：轻量化多语言视觉语言模型

Jina-VLM: Small Multilingual Vision Language Model

December 3, 2025

作者: Andreas Koukounas, Georgios Mastrapas, Florian Hönicke, Sedigheh Eslami, Guillaume Roncari, Scott Martens, Han Xiao

cs.AI

摘要

我们推出Jina-VLM——一个24亿参数的视觉语言模型，在20亿参数规模的开源多语言视觉问答模型中达到顶尖水平。该模型通过注意力池化连接器将SigLIP2视觉编码器与Qwen3语言主干网络相结合，能够以令牌高效的方式处理任意分辨率的图像。在标准视觉问答基准和多语言评估中，Jina-VLM在保持竞争力单文本性能的同时，综合表现优于同类模型。

English

We present Jina-VLM, a 2.4B parameter vision-language model that achieves state-of-the-art multilingual visual question answering among open 2B-scale VLMs. The model couples a SigLIP2 vision encoder with a Qwen3 language backbone through an attention-pooling connector that enables token-efficient processing of arbitrary-resolution images. Across standard VQA benchmarks and multilingual evaluations, Jina-VLM outperforms comparable models while preserving competitive text-only performance.