Gemini Embedding
Google's unified embedding model — topped the MTEB multilingual leaderboard at launch
Verdict
Google consolidated its embedding lineup into a single Gemini-based model that debuted at the top of the MTEB multilingual leaderboard. Matryoshka representation learning support, strong cross-lingual retrieval. Worth a serious look alongside text-embedding-3-large for new RAG builds. Gemini Embedding 2 extends the same idea to multimodal input, embedding text, images, audio, and video into one space ($0.20 per million text tokens).
Other Embedding Models
- text-embedding-3-largeProduction
OpenAI's most capable embedding model with 3072 dimensions
- Cohere Embed v3Production
Multilingual embedding model optimised for search and RAG
- Voyage AI 3Production
Specialised embeddings for code, legal, finance, and multilingual
- BGE-M3Stable
BAAI multi-granularity multilingual embedding with dense + sparse
- E5-Mistral-7BStable
Large LLM-based embedding model for maximum retrieval quality
- Jina Embeddings v3Stable
8K context multilingual embedding with task-specific LoRAs