Qwen3 Embedding 4B

Invoke Qwen3 Embedding 4B, a 4B-parameter multilingual embedding model, and turn text into vectors for retrieval, classification, and clustering.

Last updated

View as Markdown Agent setup

Qwen3 Embedding 4B is a multilingual embedding model with 4 billion parameters across 36 layers, and it supports instruction-conditioned embeddings. It reads text and returns a vector that represents its meaning, for retrieval over text and code, classification, clustering, and bitext mining. AI Inference runs it under the id Qwen/Qwen3-Embedding-4B.

Model details

The model id is the string that selects this model, and it is the first argument Azion.AI.run takes. The HuggingFace repository holds the model card.

DetailValue
Model nameQwen3 Embedding 4B
VersionOriginal
Model categoryEmbedding
Model idQwen/Qwen3-Embedding-4B
Size4B parameters
HuggingFace modelQwen/Qwen3-Embedding-4B
OpenAI-compatible endpointEmbeddings
LicenseApache 2.0

Capabilities

The model’s native vector width is 2560, and the dimensions request field selects one of the widths listed below. For dimensions, encoding_format, and the rest of the request body, refer to Model invocation.

CapabilityValue
Input dataText
Context length32k tokens
Output dimensions256, 512, 1024, 2048, 4096

Usage

A function invokes the model with Azion.AI.run, passing the id as the first argument and an OpenAI-compatible request body as the second. The examples below use that binding. To send the same body to the OpenAI-compatible HTTP endpoint instead, refer to Model invocation.

Embedding

This request carries the text in input, sets encoding_format to float, and selects a 256-dimension vector with dimensions:

TypeScript
01const modelResponse = await Azion.AI.run("Qwen/Qwen3-Embedding-4B", {
02 "input": "The food was delicious and the waiter...",
03 "encoding_format": "float",
04 "dimensions": 256
05})

The model answers with one entry in data, and the vector sits at data[0].embedding:

JSON
01{
02 "id": "embd-0123456789abcdef0123456789abcdef",
03 "object": "list",
04 "created": 1746821207,
05 "model": "Qwen/Qwen3-Embedding-4B",
06 "data": [
07 {
08 "index": 0,
09 "object": "embedding",
10 "embedding": [0.01, ..., 0.005]
11 }
12 ],
13 "usage": {
14 "prompt_tokens": 11,
15 "total_tokens": 11,
16 "completion_tokens": 0,
17 "prompt_tokens_details": null
18 }
19}
Model invocation

Every field a request body accepts, and the HTTP endpoint that takes the same body.

Vector search

Where the vectors this model returns are stored and queried.

AI Inference

The product that runs this model.

Functions

The product whose runtime holds the binding the examples on this page call.

AI Inference limits

The conditions under which Azion terminates or deprovisions a model.

Glossary

Where model id, context length, and the rest of the AI Inference vocabulary are defined.