OpenAI

OpenAI's text embedding models are industry-standard, widely-used models with excellent performance.

Overview

Models: 3 models (text-embedding-3-large, text-embedding-3-small, text-embedding-ada-002)
Key Features: Matryoshka embeddings (text-embedding-3 models), industry-standard performance
API Docs: OpenAI Embeddings

Environment Variable

export OPENAI_API_KEY="your-openai-api-key"

Supported Parameters

Parameter	Type	Required	Description
`model`	str	Yes	Model identifier (`openai:model-name`)
`input`	str \| List[str]	Yes	Text(s) to embed
`dimensions`	int	No	Output dimensions (text-embedding-3 models only)

Note: OpenAI does not support input_type. This parameter is ignored if provided.

Examples

Basic Usage

from catsu import Client

client = Client()

response = client.embed(
    "openai:text-embedding-3-small",
    "Hello, OpenAI!"
)

print(f"Dimensions: {response.dimensions}")  # 1536
print(f"Tokens: {response.usage.tokens}")

With Custom Dimensions (text-embedding-3 only)

# Reduce dimensions for faster similarity search
response = client.embed(
    "openai:text-embedding-3-small",
    ["Sample text"],
    dimensions=512  # vs default 1536
)

print(f"Dimensions: {response.dimensions}")  # 512

# Large model with custom dimensions
response = client.embed(
    "openai:text-embedding-3-large",
    ["Sample text"],
    dimensions=1024  # vs default 3072
)

Batch Processing

texts = [
    "First document",
    "Second document",
    "Third document"
]

response = client.embed(
    "openai:text-embedding-3-small",
    texts
)

print(f"Processed {len(response.embeddings)} texts")
print(f"Total tokens: {response.usage.tokens}")

Async Usage

import asyncio
from catsu import Client

async def main():
    client = Client()

    response = await client.aembed(
        "openai:text-embedding-3-small",
        "Async embedding"
    )

    print(response.embeddings)

asyncio.run(main())

Model Variants

OpenAI offers three embedding models:

text-embedding-3-large - Highest quality, 3072 dimensions
text-embedding-3-small - Balanced performance, 1536 dimensions
text-embedding-ada-002 - Legacy model, 1536 dimensions

For pricing and benchmarks, visit catsu.dev.

Special Notes

input_type is NOT supported - OpenAI ignores this parameter
Dimensions supported for text-embedding-3 models only
text-embedding-ada-002 does not support custom dimensions
Maximum 8191 tokens per input

Next Steps

Providers Overview - Compare all providers
Best Practices: Model Selection - Choose the right model

On this page