• With jina-clip-v1, users have a single embedding model that delivers state-of-the-art performance in both text-only and text-image cross-modal retrieval.
  
• Jina AI has improved on OpenAI CLIP’s performance by 165% in text-only retrieval, and 12% in image-to-image retrieval, with identical or mildly better performance in text-to-image and image-to-text tasks.
  

Highlights:

• Superior performance on all combinations of modalities, and especially large improvements in text-only embedding performance.

• Support for much longer text inputs. Jina Embeddings’ 8k token input support makes it possible to process detailed textual information and correlate it with images.

• A large net savings in space, compute, code maintenance, and complexity because this multimodal model is highly performant even in non-multimodal scenarios.