Activefastembedding Proprietary

Embeddings v3 large

by OpenAI· Released January 2024

OpenAI's most capable embedding model, offering 3072 dimensions for high-quality text embeddings. It is designed for tasks requiring nuanced semantic understanding, such as retrieval-augmented generation (RAG) and clustering. Part of the v3 family, it balances performance and cost for large-scale embedding needs.

Official Site API Docs

Input cost

$0.13 per 1M tokens

Output cost

—

Context window

8191 tokens

Max output

—

Modalities

text

License

proprietary

Capabilities

Text EmbeddingSemantic SearchClusteringClassificationRetrievalDimensionality Reduction

Best For

High-quality text embeddings for semantic search, RAG, and clustering tasks requiring 3072 dimensions.

Strengths

Highest accuracy among OpenAI embedding models
3072 dimensions for rich representations
Cost-effective at $0.13 per 1M tokens
Supports large context up to 8191 tokens

Limitations

Only supports text input (no images or audio)
Output is a fixed-size vector (no generation)
Not suitable for real-time streaming tasks
Requires API access (not open source)

Use Cases

Semantic search in enterprise document repositories

Retrieval-augmented generation (RAG) pipelines

Text clustering for topic modeling

Classification of customer feedback

Anomaly detection in text data

Recommendation systems based on content similarity

Deduplication of large text datasets

Improvements Over Previous Model

Increased dimensions from 1536 (v2) to 3072
Improved retrieval accuracy on MTEB benchmark
Lower pricing: $0.13/M tokens vs $0.20/M tokens for v2
Larger context window: 8191 tokens vs 8191 tokens (same)

Back to all models