Bootstrapping Audio Embeddings from Multimodal LLMs
Turn any multimodal LLM into a small audio embedding model that beats CLAP with 25x less data.
Latest
Identifying Embedding Models from Raw Numerical Values
A tiny transformer that fingerprints embedding models by reading raw numerical digits. No feature engineering.
jina-embeddings-v5-text: New SOTA Small Multilingual Embeddings
Two sub-1B multilingual embeddings with best-in-class performance, available on Elastic Inference Service, Llama.cpp and MLX.
Jina-VLM: Small Multilingual Vision Language Model
New 2B vision language model achieves SOTA on multilingual VQA, no catastrophic forgetting on text-only tasks.
Jina Reranker v3: 0.6B Listwise Reranker for SOTA Multilingual Retrieval
New 0.6B-parameter listwise reranker that considers the query and all candidate documents in a single context window.
Embeddings Are AI’s Red-Headed Stepchild
Embedding models aren't the most glamorous aspect of the AI industry, but image generators and chatbots couldn't exist without them.
Multimodal Embeddings in Llama.cpp and GGUF
We brought multimodal embeddings to llama.cpp and GGUF, and uncovered a few surprising issues along the way.
Jina Code Embeddings: SOTA Code Retrieval at 0.5B and 1.5B
Code generation LLMs → code embeddings: 0.5B/1.5B models achieve SOTA performance across 25 code retrieval benchmarks.
Agentic Workflow with Jina Remote MCP Server
Jina MCP streamlines agent development by connecting our APIs to any LLM, reducing custom code and improving reliability of the workflow.
Optimizing GGUFs for Decoder-Only Embedding Models
4000 tokens/sec for a 3B-parameter embedding model on L4 GPU is probably as fast as you'll get with llama.cpp. Or is it?
What We Learned at SIGIR 2025
Sharing what we saw and learned at SIGIR 2025, feat. CLIP-AdaM, RE-AdaptIR and evaluations for LLM-based retrieval systems.
How Image Resolution Impacts Visual Document Retrieval
Image resolution is crucial for embedding visually rich documents. Too small and models miss key details; too large and they can't connect the parts.
Press
JinaVDR: New Visual Document Retrieval Benchmark with 95 Tasks in 20 Languages
JinaVDR is a new benchmark spanning 95 tasks across 20 languages for visual document retrieval, soon on MTEB.
Tech Blog
Submodular Optimization for Text Selection, Passage Reranking & Context Engineering
While others rely on prompt tuning and hope for the best, you should learn submodular optimization that provides a principled framework with theoretical guarantees for better context engineering.
Tech Blog
Submodular Optimization for Diverse Query Generation in DeepResearch
Many know the importance of query diversity in DeepResearch, but few know how to solve it rigorously via submodular optimization.
Tech Blog
Quantization-Aware Training of jina-embeddings-v4
Quantization gives smaller embeddings. We show you fine-tuned quantization gives you even lossless embeddings.
Press
Jina Embeddings v4: Universal Embeddings for Multimodal Multilingual Retrieval
Jina Embeddings v4 is a 3.8 billion parameter universal embedding model for multimodal and multilingual retrieval that supports both single-vector and multi-vector embedding outputs.
Tech Blog
Correlations: Vibe-Testing Embeddings in GUI
As serious as we are about MTEB, we also love vibe-testing. Correlations is a simple GUI we use for validating citations in DeepSearch, debugging late chunking, and vibe-testing embeddings. Now it's open-source.
Events
What We Learned at ICLR2025
We collect some most interesting papers in ICLR 2025, featuring TIPS, FlexPrefill, Zero-Shot Rerankers, SVD-LLM, Hymba etc.
Tech Blog
Fair Scoring for Multimodal Documents with jina-reranker-m0
Text similarity: 0.7. Image similarity: 0.5. Which document is more relevant? You literally cannot tell—and that's the core problem breaking multimodal search. We solve it with unified reranking.
Tech Blog
Model Soup’s Recipe for Embeddings
Boost robustness and performance with model soups: averaging weights. No extra cost, better results.
Tech Blog
On the Size Bias of Text Embeddings and Its Impact in Search
Size bias refers to how the length of text inputs affects similarity, regardless of semantic relevance. It explains why search systems sometimes return long, barely-relevant documents instead of shorter, more precise matches to your query.
Press
jina-reranker-m0: Multilingual Multimodal Document Reranker
Introducing jina-reranker-m0, our new multilingual multimodal reranker for retrieving visual documents, with SOTA performance on multilingual long documents and code searching tasks.
Tech Blog
Using DeepSeek R1 Reasoning Model in DeepSearch
Standard LLM or reasoning model, which is better for DeepSearch? In this post, we explored using DeepSeek-R1 in the DeepSearch implementation for choosing the next action.