OpenAI Text Embeddings
OpenAIOpenAI's general-purpose text embedding models, text-embedding-3-small and text-embedding-3-large, reached only through the /v1/embeddings endpoint. Both accept 8,192 tokens, default to 1,536 and 3,072 dimensions, and were trained with Matryoshka-style shortening so a caller can trade dimensions for storage through the `dimensions` parameter.
OpenAI's embeddings guide still lists only text-embedding-3-small, text-embedding-3-large and the legacy text-embedding-ada-002; there is no text-embedding-4.
Openness
1 high confidence- weights
- closed(no checkpoint distributed
- data
- closed(training corpus neither released nor described)
- code
- closed(no training or inference implementation
- license
- proprietary(hosted API under OpenAI business terms
API-only: the guide documents an endpoint, a dimensions parameter and a price, but nothing that can be downloaded, which settles the score ahead of any license question.
- https://developers.openai.com/api/docs/guides/embeddings recorded 2026-09-11
Documents text-embedding-3-small and text-embedding-3-large as endpoint calls with an 8,192-token limit and a `dimensions` parameter; the page offers no weight download, no corpus description and no training code, and names no successor model.
Adoption
4 medium confidenceOpenAI publishes no download or call-volume figure for its embedding models, so there is no artifact to count. Adoption is read from distribution and standing instead: Microsoft resells both SKUs through Azure AI Foundry, and rival vendors benchmark against OpenAI v3 Large as their reference point.
- https://blog.voyageai.com/2026/01/15/voyage-4/ recorded 2026-09-11
Voyage's RTEB evaluation picks 'OpenAI v3 Large' as one of three external reference models to measure against, alongside Gemini Embedding 001 and Cohere Embed v4.
- https://learn.microsoft.com/en-us/azure/ai-foundry/openai/concepts/models recorded 2026-09-11
Azure AI Foundry's models-sold-directly-by-Azure list carries text-embedding-3-large and text-embedding-3-small, so the models are resold by Microsoft as well as served by OpenAI.
Capability
3 high confidenceText-embedding-3-large and -small are a workhorse generation: their MTEB scores are credible and the models see enormous everyday use, but they are short-context at 8,192 tokens, text-only, and post the largest measured gap in Voyage's own RTEB comparison. That is a description of scope rather than a demotion.
- https://blog.voyageai.com/2026/01/15/voyage-4/ recorded 2026-09-11
'voyage-4-large is the top-performing model, surpassing voyage-4, voyage-4-lite, Gemini Embedding 001, Cohere Embed v4, and OpenAI v3 Large by an average of 1.87%, 4.80%, 3.87%, 8.20%, and 14.05%, respectively' over all 29 RTEB datasets.
- https://developers.openai.com/api/docs/guides/embeddings recorded 2026-09-11
Model table: text-embedding-3-small 62.3%, text-embedding-3-large 64.6%, text-embedding-ada-002 61.0% on 'Performance on MTEB eval', max input 8192 for all three.
Verified 2026-09-11