Introducing voyage-3.5 and voyage-3.5-lite: Improved Quality for a New Retrieval Frontier

Voyage introduces its new embedding models, voyage-3.5 and voyage-3.5-lite, offering enhanced retrieval quality at the same cost as their predecessors, voyage-3 and voyage-3-lite. These models support embeddings in multiple dimensions, utilizing Matryoshka learning and quantization-aware training, which enable a range of quantization options. The new models outperform OpenAI-v3-large in retrieval quality while significantly reducing vector database costs. Voyage-3.5 and voyage-3.5-lite, evaluated across 100 datasets in diverse domains including technical documentation, code, law, and finance, show superior performance in retrieving documents based on cosine similarities. Additionally, the models are available through the Voyage AI APIs and the private preview of automated text embedding in Atlas Vector Search, providing a cost-effective and high-performance solution for various applications.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.