Code beginner · 3 min read

How to generate embeddings with Python

Direct answer

Use the OpenAI SDK's embeddings.create method with a model like text-embedding-3-small to generate vector embeddings from text in Python.

Setup

Install

bash

pip install openai

Env vars

OPENAI_API_KEY

Imports

python

import os
from openai import OpenAI

Examples

inHello world

out[0.0123, -0.0456, 0.0789, ...] # list of floats representing the embedding vector

inHow to generate embeddings with Python?

out[0.0345, -0.0678, 0.0234, ...] # embedding vector for the question text

out[] # empty input returns an empty or default embedding vector

Integration steps

Install the OpenAI Python SDK and set the OPENAI_API_KEY environment variable.
Import the OpenAI client and initialize it with the API key from os.environ.
Call the embeddings.create method with the desired model and input text.
Extract the embedding vector from the response's data[0].embedding field.
Use or store the embedding vector for downstream tasks like search or classification.

Full code

python

import os
from openai import OpenAI

# Initialize OpenAI client with API key from environment
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Text to embed
text_to_embed = "How to generate embeddings with Python?"

# Create embeddings
response = client.embeddings.create(
    model="text-embedding-3-small",
    input=text_to_embed
)

# Extract embedding vector
embedding_vector = response.data[0].embedding

print(f"Embedding vector for input: {embedding_vector[:5]}... (truncated)")

API trace

Request

json

{"model": "text-embedding-3-small", "input": "How to generate embeddings with Python?"}

Response

json

{"data": [{"embedding": [0.02345, -0.01234, 0.04567, ...]}], "usage": {"prompt_tokens": 8, "total_tokens": 8}}

Extractresponse.data[0].embedding

Variants

Batch Embeddings Generation ›

python

import os
from openai import OpenAI

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

texts = ["Hello world", "Generate embeddings with Python", "OpenAI embeddings"]

response = client.embeddings.create(
    model="text-embedding-3-small",
    input=texts
)

embeddings = [item.embedding for item in response.data]
for i, emb in enumerate(embeddings):
    print(f"Embedding {i} (truncated): {emb[:5]}...")

Async Embeddings Generation ›

python

import os
import asyncio
from openai import OpenAI

async def generate_embedding(text: str):
    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
    response = await client.embeddings.acreate(
        model="text-embedding-3-small",
        input=text
    )
    return response.data[0].embedding

async def main():
    embedding = await generate_embedding("Async embeddings with Python")
    print(f"Async embedding (truncated): {embedding[:5]}...")

asyncio.run(main())

Alternative Model for Larger Embeddings ›

python

import os
from openai import OpenAI

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

response = client.embeddings.create(
    model="text-embedding-3-large",
    input="Generate higher dimensional embeddings"
)

embedding = response.data[0].embedding
print(f"Large embedding vector (truncated): {embedding[:5]}...")

Performance

Latency~300-600ms per embedding request for single input on text-embedding-3-small

Cost~$0.0004 per 1,000 tokens for text-embedding-3-small

Rate limitsTier 1: 60 RPM / 60,000 TPM typical for embeddings endpoint

Minimize input length to reduce token usage and cost.
Batch inputs to amortize overhead per request.
Use smaller embedding models for less critical tasks.

Approach	Latency	Cost/call	Best for
Single input embedding	~300-600ms	~$0.0004	Quick embedding of one text
Batch embedding	~500-1000ms	~$0.0004 per 1k tokens total	Efficient multi-text embedding
Async embedding	Varies, concurrent	Same as sync	Embedding in async apps
Large model embedding	~700-1200ms	~$0.001	High-quality embeddings

✓

Quick tip

Always batch multiple texts in one embeddings.create call to optimize latency and cost.

⚠

Common mistake

Passing an empty string or non-string input to embeddings.create often results in errors or meaningless embeddings.

Verified 2026-04 · text-embedding-3-small, text-embedding-3-large

Verify ↗