voyage-3-lite
Voyage 3 Lite is the smaller text embedding release launched with voyage-3. It targets lower-latency retrieval.
Model design
This version accepts 32,000-token inputs and returns 512-dimensional vectors. Its fixed output differs from the selectable dimensions introduced in later Lite models.voyage-3 and voyage-3-lite
Practical use
Practical guidance. Smaller vectors reduce the amount of vector data stored per passage. Compare retrieval at the same passage boundaries and candidate count so a storage change is not confused with a change in chunking.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | API |
| Access | Earlier Voyage API modelVoyage AI text embedding catalog and earlier models |
Related models
- voyage-4-lite
Voyage AI · Voyage 4
Smaller hosted text encoder in the shared Voyage 4 family.
Launch date API - voyage-3.5-lite
Voyage AI · Voyage 3.5
Smaller 3.5-series embeddings with flexible vector sizes.
Launch date API Third-generation general text embeddings with a 32,000-token input limit.
Launch date API