voyage-multimodal-3.5

About this model

Multimodal embedding models are neural networks that transform multiple modalities, such as text and images, into numerical vectors. They are a crucial building block for semantic search/retrieval systems and retrieval-augmented generation (RAG) and are responsible for the retrieval quality. voyage-multimodal-3.5 is a state-of-the-art multimodal embedding model capable of vectorizing not only text, images, and video individually, but also content that interleaves all three modalities. It delivers excellent performance for mixed-modality searches involving text and visual content such as PDF screenshots, figures, tables, videos, and more. Enabled by Matryoshka learning and quantization-aware training, voyage-multimodal-3.5 supports embeddings in 2048, 1024, 512, and 256 dimensions, with multiple quantization options. Learn more about voyage-multimodal-3.5 here: https://blog.voyageai.com/2026/01/15/voyage-multimodal-3-5

Key model capabilities

• State-of-the-art multimodal embedding model capable of vectorizing not only text, images, and video individually, but also content that interleaves all three modalities. It delivers excellent performance for mixed-modality searches involving text and visual content such as PDF screenshots, figures, tables, videos, and more.
• Supports embeddings of 2048, 1024, 512, and 256 dimensions and offers multiple embedding quantization, including float (32-bit floating point), int8 (8-bit signed integer), uint8 (8-bit unsigned integer), binary (bit-packed int8), and ubinary (bit-packed uint8).
• 32K token context length.

Usage

The deployed Azure AI Foundry endpoint exposes the Voyage multimodal inference API. Authenticate with your Azure ML endpoint key or a bearer token issued for the workspace.

Generate Embeddings

curl <AZUREML_ENDPOINT_URL>/v1/multimodalembeddings \
  -X POST \
  -H "Authorization: Bearer <AZUREML_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{
    "model":"voyage-multimodal-3.5",
    "input_type":"document",
    "inputs":[
      {"content":[
        {"type":"text","text":"A penguin on an iceberg"},
        {"type":"image_url","image_url":"https://example.com/penguin.jpg"}
      ]}
    ]
  }'

Each entry in inputs may interleave text, image, and video parts. Valid content[].type values are text, image_url, image_base64, video_url, and video_base64. Video parts require voyage-multimodal-3.5.

Supported Parameters

• inputs (array, required): 1–1000 items, each with a content array of text/image/video parts.
• model (string, required): voyage-multimodal-3.5.
• input_type (string, optional): query or document.
• output_dimension (int, optional): One of 2048, 1024, 512, 256.
• output_dtype (string, optional): float, int8, uint8, binary, or ubinary.
• truncation (bool, optional): Defaults to true.
• encoding_format (string, optional): Set to base64 to receive embeddings as base64 strings.

See the full API reference at https://docs.voyageai.com/reference/multimodal-embeddings-api .

Response

{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "index": 0,
      "embedding": [0.00068755, 0.03410244, -0.02404458, 0.04494607]
    }
  ],
  "model": "voyage-multimodal-3.5",
  "usage": {
    "total_tokens": 256,
    "text_tokens": 6,
    "image_pixels": 250000,
    "video_pixels": 0
  }
}

The embedding array contains the full vector at the requested output_dimension (shown truncated above). usage reports text-token, image-pixel, and video-pixel consumption separately.