text-embedding-3-large
OpenAI
OpenAI's largest text embedding model (January 2024), producing up to 3,072-dimensional vectors for retrieval, RAG, clustering, and classification. It supports shortening dimensions via the API for storage/latency trade-offs and remains a strong general-purpose retrieval baseline.
Strengths
- Strong general-purpose retrieval quality (~64.6 MTEB)
- Up to 3,072-dimensional embeddings
- Shortenable dimensions for storage/latency trade-offs
- Simple, reliable API integration
- Widely supported across vector databases
Caveats
- Proprietary - API only (no self-hosting)
- Not updated since Jan 2024; newer models lead benchmarks
- English-centric relative to top multilingual models
- Vendor
- OpenAI
- Context Window
- 8,191 tokens
- License
- Proprietary
- Release Date
- 2024-01-25
- Modalities
- text
Capabilities
Vision
Audio
Video
Tool Use
Resources
Pricing
$0.13 per 1M input tokens via the OpenAI API. Dimensions are shortenable (e.g. 256/1024/3072) to trade quality for storage.