Voyage 4 Large
Available nowA retrieval-focused choice for finding the right evidence in demanding knowledge collections.
What it does
Voyage 4 Large is positioned by Voyage AI as its strongest general-purpose and multilingual retrieval model in the Voyage 4 series. Its job is to connect a question with the passages most likely to contain useful information.
The family supports long source context and four vector sizes. Voyage documents compatible vector spaces across the general Voyage 4 series, giving teams options when balancing query quality and indexing cost. That compatibility should still be validated with matching dimensions and retrieval settings in your application.
Language coverage
Voyage describes the current general retrieval family as multilingual. The model guide does not publish a complete named-language guarantee; treat it as a broad candidate and evaluate your collection.
Official language coverage ↗Where it helps
Enterprise knowledge
Find the right evidence in large internal document collections.
Customer-facing assistants
Retrieve relevant passages for questions that simpler search misses.
Research archives
Connect detailed questions with useful multilingual reference material.
When to choose it
A candidate when retrieval accuracy is a priority and search failures are expensive. Evaluate it against Voyage 4 and 4 Lite on the same real questions before deciding how much quality you need.
embeddings.app exposes text float vectors and its own byte limits; the provider’s full token window does not equal the service’s maximum input size.
Official sources
Paraphrased from the provider’s documentation and original model information. Selection advice is our editorial guidance.
Voyage AI · Model guide ↗Reviewed October 3, 2026