text-embedding-3
A text embedding model series released by OpenAI, employing Matryoshka Representation Learning to support variable-dimension embeddings. It is suitable for tasks including semantic search and text similarity measurement.
Timeline (last 90 days)
Verdandi uses the GPT-6-luna model, and ACU uses the GPT-5.5 and text-embedding-3 models.
OpenAI's text-embedding-3 series employs a technology similar to MRL.
OpenAI's text-embedding-3 series natively supports the Matryoshka feature, allowing users to specify the output dimension via an API parameter.
Current mainstream embedding models perform significantly worse at handling numbers and precise data than at handling concepts and semantics.
Mainstream text embedding models are typically trained based on contrastive learning, mapping semantically similar text pairs to nearby positions in vector space and pushing unrelated text pairs further apart.
All Facts (6)
BGE和text-embedding-3是当前可选的向量嵌入（Embedding）模型
90%UnverifiedVerdandi uses the GPT-6-luna model, and ACU uses the GPT-5.5 and text-embedding-3 models.
50%UnverifiedOpenAI's text-embedding-3 series employs a technology similar to MRL.
50%UnverifiedOpenAI's text-embedding-3 series natively supports the Matryoshka feature, allowing users to specify the output dimension via an API parameter.
50%UnverifiedCurrent mainstream embedding models perform significantly worse at handling numbers and precise data than at handling concepts and semantics.
50%UnverifiedMainstream text embedding models are typically trained based on contrastive learning, mapping semantically similar text pairs to nearby positions in vector space and pushing unrelated text pairs further apart.
50%