Activefastembedding Proprietary

Embeddings v3 small

by OpenAI· Released January 2024

OpenAI's most efficient embedding model, optimized for speed and cost while maintaining strong performance. It is part of the third-generation embedding family, offering a smaller size for lower latency and reduced computational requirements.

Official Site API Docs

Input cost

$0.02 per 1M tokens

Output cost

—

Context window

8191 tokens

Max output

—

Modalities

text

License

proprietary

Capabilities

Text EmbeddingSemantic SearchClusteringClassificationCosine Similarity

Best For

High-throughput, cost-sensitive embedding tasks where speed is prioritized over maximum accuracy.

Strengths

Fast inference speed
Low cost per token
Strong performance on retrieval and clustering benchmarks
Efficient for large-scale applications

Limitations

Lower accuracy compared to larger embedding models like text-embedding-3-large
Not suitable for tasks requiring very high precision
Limited to text input only

Use Cases

Semantic search for large document corpora

Text clustering for content organization

Recommendation systems based on text similarity

Anomaly detection in text data

Customer support ticket routing

Duplicate detection in databases

Real-time embedding for chatbots

Improvements Over Previous Model

Reduced pricing from $0.10 per 1M tokens (text-embedding-ada-002) to $0.02 per 1M tokens
Improved retrieval performance on MTEB benchmark
Smaller model size for faster inference
Supports 8191 token context window vs 8191 (same as predecessor)

Back to all models