Plain-English Summary
A compact, inexpensive embedding model for search and retrieval at scale.
text-embedding-3-small
A compact, inexpensive embedding model for search and retrieval at scale.
Embedding 3
Embedding
GA
Open weights: No
API: Yes
Technical Notes
Proprietary text embedding model with up to 1,536 dimensions and an 8,191-token input limit.
Best For
Cost-effective RAG, deduplication, search, recommendations
Watch Out For
Lower retrieval quality than the large model on demanding datasets.
Model Specs
- Maker
- OpenAI
- Country
- USA
- Family
- Embedding 3
- Type
- Embedding
- Release date
- 2024-01-25
- Date confidence
- Exact
- Status
- GA
- Context window
- 8,191 tokens
- Max output
- Not listed
- Parameters
- Not disclosed
- Active parameters
- Not disclosed
- Architecture
- Proprietary embedding transformer
- Reasoning
- No / not emphasized
- Tool calling
- No / not documented
- Structured output
- No / not documented
- API available
- Yes
- Open weights
- No
- Self-hostable
- No
- License
- Proprietary
- Inputs
- Text
- Outputs
- Embedding vector
- Approx. price
- $0.02 per 1M input tokens
- How to access
- Provider website or API
- Model / API ID
- text-embedding-3-small
- Knowledge cutoff
- Not publicly disclosed
- Fine-tuning
- Selectable embedding dimensions
- Completeness
- 86 %
- Verified
- Jul 18, 2026
- Snapshot date
- Jul 18, 2026
- Source
- Official source