How to beginner · 4 min read

How to use BGE embeddings

Q: How to use BGE embeddings

Use the BGE embeddings model via the OpenAI API by calling client.embeddings.create with model="bge-small" or "bge-base". This returns a vector representation of input text useful for semantic search, clustering, or similarity tasks.

Quick answer

Use the BGE embeddings model via the OpenAI API by calling client.embeddings.create with model="bge-small" or "bge-base". This returns a vector representation of input text useful for semantic search, clustering, or similarity tasks.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0

Setup

Install the openai Python package and set your OpenAI API key as an environment variable for authentication.

bash

pip install openai>=1.0

Step by step

Use the OpenAI Python SDK to generate BGE embeddings for your text input. The example below shows how to create embeddings with the bge-small model and print the resulting vector length.

python

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

response = client.embeddings.create(
    model="bge-small",
    input="OpenAI's BGE embeddings provide powerful text representations."
)

embedding_vector = response.data[0].embedding
print(f"Embedding vector length: {len(embedding_vector)}")

output

Embedding vector length: 768

Common variations

Use bge-base for higher-dimensional embeddings with potentially better accuracy.
Pass a list of strings to input to batch embed multiple texts in one request.
Use async calls with asyncio and the OpenAI async client for concurrency.

python

import os
import asyncio
from openai import OpenAI

async def async_embed():
    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
    response = await client.embeddings.acreate(
        model="bge-base",
        input=["First text", "Second text"]
    )
    for i, data in enumerate(response.data):
        print(f"Embedding {i} length: {len(data.embedding)}")

asyncio.run(async_embed())

output

Embedding 0 length: 1024
Embedding 1 length: 1024

Troubleshooting

If you get an authentication error, verify your OPENAI_API_KEY environment variable is set correctly.
For rate limit errors, implement exponential backoff retries.
If embeddings seem off, check you are using the correct model name like bge-small or bge-base.

✅

Key Takeaways

Use bge-small or bge-base models with OpenAI's embeddings API for BGE embeddings.
Batch multiple texts in one request by passing a list to input for efficiency.
Async API calls improve throughput when embedding many texts concurrently.
Always set your API key in OPENAI_API_KEY environment variable to avoid auth errors.
Check model names carefully to avoid unexpected results or errors.

Verified 2026-04 · bge-small, bge-base

Verify ↗