Voyage AI / voyageai/voyage-context-4

Voyage Context 4 - access through LLMTR

Voyage Context 4 vectorizes a document's chunks by seeing all of them at once rather than one at a time. Each chunk's vector therefore carries its position within the document, so phrases such as "this ratio" or "the said party" can resolve against a definition that appears in a different chunk. That context is lost under classic chunking and recovering it measurably improves retrieval over long technical documents and contracts. Unlike the other embedding models, input is supplied as a nested array: the outer list holds documents and each inner list holds that document's chunks.

Technical specifications

Canonical IDvoyageai/voyage-context-4
ProviderVoyage AI
Context window32,000 tokens
OperationsEMBEDDINGS
Modalitiestext, embedding

Pricing

An 8% platform margin applies to credit top-ups; model usage prices are not separately marked up.

OperationMetricUnitPrice
EMBEDDINGSINPUT_TEXTPER_1M_TOKENS$0.180000

Example usage

With existing OpenAI SDK flows, change only the base URL and model identifier.

curl https://llmtr.com/v1/chat/completions \
  -H "Authorization: Bearer llmtr-your_key" \
  -H "Content-Type: application/json" \
  -d '{"model":"voyageai/voyage-context-4","messages":[{"role":"user","content":"Hello"}]}'

Related models