Jina-VLM: Small Multilingual Vision Language Model

Jina AI · Product launch · December 4, 2025

Jina AI released Jina-VLM, a 2.4B parameter vision-language model designed for multilingual visual question answering. The model utilizes an attention-pooling connector to process images efficiently while maintaining strong performance on text-only tasks.

Read the press release

AntSpace’s summary of Jina AI’s own announcement: read the original.

More updates from Jina AI