Text Embedding 3 Small

• 🔧 OpenAI text embedding model for search, clustering and retrieval
• 🆕 Improved, more performant successor to the ada embedding model
• 📏 Returns 1536-dimension vectors by default
• ⚡ Smaller counterpart to [[sibling:text-embedding-3-large|Text Embedding 3 Large]]
• 🎯 Powers semantic search, recommendations and anomaly detection
• 📚 Common backbone for retrieval-augmented generation pipelines
• 🏢 Released alongside the larger v3 embedding model
• 🌐 Available through the standard OpenAI embeddings API endpoint

OpenAI

OpenAI is an American artificial intelligence research organization headquartered in San Francisco, structured as both a for-profit public benefit corporation and a nonprofit foundation. The lab developed the GPT family of large language models, the DALL-E…

Text Embedding 3 Small is OpenAI's compact text-embedding model, designed to convert text into numerical vectors that capture semantic relatedness. These embeddings underpin tasks such as semantic search, clustering, recommendations, classification and anomaly detection, and are a standard building block for retrieval-augmented generation systems that inject external context into a language model's prompt.

According to OpenAI's own documentation, text-embedding-3-small is "our improved, more performant version of our ada embedding model," positioning it as a direct successor to the earlier ada generation rather than a sibling tweak. It is offered as the smaller counterpart to Text Embedding 3 Large, the higher-capacity model released the same day in OpenAI's v3 embedding family.

By default the model produces vectors of 1536 dimensions, which downstream tools store and compare using similarity measures such as cosine similarity.

Integration is straightforward: the model is called through the standard OpenAI embeddings endpoint by passing input text and the model name, returning an array of floating-point values that represent the input in vector space.