Gemini Embedding: Generalizable Embeddings from Gemini
Jinhyuk Lee, Feiyang Chen, Sahil Dua, Daniel Cer, Madhuri Shanbhogue, Iftekhar Naim, Gustavo Hernández Ábrego, Zhe Li, Kaifeng Chen, Henrique Schechter Vera, Xiaoqi Ren, Shanfeng Zhang, Daniel Salz, Michael Boratko, Jay Han, Blair Chen, Shuo Huang, Vikram Rao, Paul Suganthan, Feng Han, Andreas Doumanoglou, Nithi Gupta
2025-03-12
Summary
This paper talks about Gemini Embedding, a smart AI tool that turns text into math-like codes (embeddings) to help computers understand and compare words, sentences, or code across different languages and tasks.
What's the problem?
Existing AI models struggle to handle text in many languages or specialized formats (like code) at once, often needing separate tools for each task or language.
What's the solution?
Gemini Embedding uses Google’s powerful Gemini AI to create universal codes for text, trained on a mix of real and AI-generated examples to work across 250+ languages and tasks like translation or code analysis.
Why it matters?
This lets apps like search engines or translation tools work better across languages and tasks, helping people communicate and find information globally without needing multiple AI systems.
Abstract
In this report, we introduce Gemini Embedding, a state-of-the-art embedding model leveraging the power of Gemini, Google's most capable large language model. Capitalizing on Gemini's inherent multilingual and code understanding capabilities, Gemini Embedding produces highly generalizable embeddings for text spanning numerous languages and textual modalities. The representations generated by Gemini Embedding can be precomputed and applied to a variety of downstream tasks including classification, similarity, clustering, ranking, and retrieval. Evaluated on the Massive Multilingual Text Embedding Benchmark (MMTEB), which includes over one hundred tasks across 250+ languages, Gemini Embedding substantially outperforms prior state-of-the-art models, demonstrating considerable improvements in embedding quality. Achieving state-of-the-art performance across MMTEB's multilingual, English, and code benchmarks, our unified model demonstrates strong capabilities across a broad selection of tasks and surpasses specialized domain-specific models.