Romain Huet· Updated
OpenAI confirmed that text-embedding-3-small is not being deprecated, correcting a previous communication error that suggested the model would be retired. This clarification prevents a massive re-indexing requirement for developers who rely on the model for their vector databases and RAG pipelines.
text-embedding-3-small is not being deprecated, reversing a previous notice that suggested the model would be retired. Romain Huet, Head of Developer Experience, clarified that the communication was a mistake. This model is a primary tool for generating embeddings (numerical representations of text used for semantic search).The update follows a period of model pruning, including when OpenAI retired early GPT-5 models from its interface. Unlike chat models, retiring an embedding model forces developers to re-index their entire knowledge base. This clarification provides critical stability for teams building Retrieval-Augmented Generation (RAG) systems (grounding AI in external data).
You do not need to migrate vector stores or change API configurations for text-embedding-3-small. OpenAI will email users to formally clarify the model's status and is investigating where the incorrect information originated. The model remains available for all projects via standard API endpoints.
@jeffreyhuber Thank you for flagging this, Jeff. This was a mistake: we are not deprecating text-embedding-3-small. We’re looking into where this came from now, and we’ll also email users to clarify. Sorry for the confusion!
Still wondering? A few quick answers below.
Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →
