Qwen3.7 Text Embedding

It is a multilingual unified text embedding model trained by Tongyi Laboratory based on Qwen3.7. Compared to the text-embedding-v4 version, it significantly improves performance in text retrieval, clustering, and classification; its performance improves by 20% on evaluation tasks such as MTEB multilingual, Chinese-English, and Code retrieval; it supports user-customized vector dimensions ranging from 256 to 2560 dimensions.