voyage-4-large

Embedding modelCard summarized

The best general-purpose and multilingual retrieval quality.

Source: Voyage 4 model family launch post (no model card published)

Use it via API

$0.12 / 1M tok

Code example Compare

Where to call it

Voyage AI APIvoyage-4-large

Supports

Batch API

Full feature support

Good for

General-purpose and multilingual retrieval at the highest Voyage 4 quality
Indexing documents once with large while serving queries from a cheaper Voyage 4 model (asymmetric retrieval)
Storage savings via 256–2048 dims and int8/binary outputs

Watch out for

Over-length text is truncated silently
Smaller batches than other Voyage 4 models
Language coverage not listed

Overview

Voyage AI · 01Large

voyage-4-large

The best general-purpose and multilingual retrieval quality.

Typical price

$0.12 / 1M tok

Text input

Dimensions: 1024 default
Max input: 32,000 tokens
Modes: text embeddings, multilingual retrieval
Released: 15 Jan 2026

voyage-4-large is the top model of the Voyage 4 text embedding family (voyage-4-large, voyage-4, voyage-4-lite and open-weights voyage-4-nano). Voyage calls it the first production-grade embedding model to use a mixture-of-experts architecture. All Voyage 4 models write into one shared embedding space, so documents indexed with voyage-4-large can be searched with queries from a smaller model. launch post

voyage-4-large

Typical price: $0.12 / 1M tok
Dimensions: 1024 default (can be set to 256, 512 or 2048)
Max input: 32,000 tokens
Languages: Multilingual (languages not listed)
Weights: Closed
Released: 15 Jan 2026

text embeddings
multilingual retrieval