text-embedding-v4

Text-Embedding-V4 is an advanced multilingual text embedding model developed by Alibaba Cloud's Tongyi Lab and built on the Qwen3 architecture. It transforms natural language text into high-quality semantic vector representations and supports over 100 languages, including multiple programming languages. The model delivers strong performance in semantic retrieval, text similarity, classification, and clustering. With configurable embedding dimensions ranging from 64 to 2048, it is ideal for knowledge retrieval, RAG applications, recommendation systems, and enterprise semantic search.