No longer supported

This model is legacy (released December 2022) and is no longer actively maintained or recommended by OpenAI. Consider using their current flagship model instead.

legacylegacyembedding Proprietary

Embeddings v2 ada

by OpenAI· Released December 2022· Cutoff September 2021

OpenAI's second-generation embedding model, text-embedding-ada-002, converts text into 1536-dimensional vectors for semantic similarity and search. It replaced the earlier Ada v1 models and was the default embedding model until the release of v3 models in early 2024. It is now deprecated in favor of newer, more capable embedding models.

Official Site API Docs

Input cost

$0.10 per 1M tokens

Output cost

—

Context window

8191 tokens

Max output

—

Modalities

text

License

proprietary

Capabilities

Semantic SearchClusteringRecommendationAnomaly DetectionClassification

Best For

Generating high-quality text embeddings for semantic search, clustering, and classification tasks.

Strengths

High-quality 1536-dimensional embeddings
Cost-effective at $0.10/1M tokens
Good performance on semantic similarity benchmarks

Limitations

Legacy model, deprecated in favor of v3 models
Only supports text input
No longer receiving updates or improvements

Use Cases

Semantic search for documents

Text clustering for topic modeling

Recommendation systems based on content similarity

Anomaly detection in text data

Classification of text into categories

Duplicate detection

Retrieval-augmented generation (RAG) pipelines

Improvements Over Previous Model

Replaced earlier Ada v1 embedding models with higher quality embeddings
Increased context window from 2048 to 8191 tokens
Reduced pricing from $0.20 to $0.10 per 1M tokens
Improved performance on semantic similarity benchmarks (e.g., STS) compared to v1

Back to all models