Voyage multimodal 3
Multimodal embedding model that vectorizes interleaved text and images through a unified transformer encoder. Supports screenshots of PDFs, slides, tables, and figures without complex document parsing. Unlike CLIP-based models, eliminates the modality gap, enabling accurate mixed-modality retrieval across text and visual content.
- Type
- Multimodal
- Weights
- Closed
- Likes
- 16
- License
- mit
Model hub
voyageai/voyage-multimodal-3
- Repo
- voyageai/voyage-multimodal-3
- Access
- Public
More from Voyage AI
11
-
Rerank 2 lite
Text
rerank-2-lite is a lightweight multilingual cross-encoder reranker from Voyage AI. It scores and reorders can…
-
rerank 2.5
Text
A reranking model that reorders search results to improve retrieval accuracy, with instruction-following capa…
-
Rerank 2.5 lite
Text
A latency-optimized reranker that refines retrieval results with instruction-following capabilities. Supports…
-
Voyage
Text
voyage is a text embedding model that converts text into dense vector representations for semantic retrieval.…
-
Voyage 01
Text
voyage-01 is a text embedding model by Voyage AI, designed for retrieval-augmented generation (RAG) and seman…
-
Voyage 3 large
Text
Text embedding model optimized for general-purpose and multilingual retrieval across law, finance, code, and…
-
Voyage 3 lite
Text
Lightweight general-purpose text embedding model optimized for latency and cost. Generates 512-dimensional ve…
-
Voyage 4
Text
Text embedding model optimized for general-purpose and multilingual retrieval quality. Supports a 32,000-toke…
Pricing
1
About Voyage AI
1
Voyage AI provides best-in-class embedding models and rerankers for search and retrieval over unstructured data, used to power retrieval-augmented generation (RAG) and AI applications. It offers general-purpose, domain-specific (finance, legal, code) and company-specific fine-tuned models. Founded in 2023 and based in Palo Alto, the company was acquired by MongoDB, Inc. in February 2025 and now operates as a MongoDB subsidiary.