Qwen3-Embedding-4B

Name in the API

Qwen/Qwen3-Embedding-4B

Base model

Qwen/Qwen3-Embedding-4B

Type

Embeddings

Input → output

text → vector with 2,560 dimensions

Context length

40,960 tokens per input

Availability

Dedicated

Hardware

A 24 GB slice of a GPU per instance

Endpoint

/v1/embeddings

Example

Embeddings and text similarity

Notes

• For search, the model authors recommend adding a one-sentence task description to the queries (not to the documents), for example Instruct: Given a web search query, retrieve relevant passages that answer the query\nQuery: <your query>.