---
title: "Voyage AI's voyage-4 generation cut price only at its flagship tier; the mid and budget tiers stayed exactly flat"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified (docs.voyageai.com/docs/pricing and blog.voyageai.com read directly by the capture at capture time, 2026-07-15; queen independent re-fetch not attempted in this headless promotion pass)"
source_url: ["https://docs.voyageai.com/docs/pricing","https://blog.voyageai.com/2026/01/15/new-models-and-expanded-availability/"]
source_author: "Voyage AI (official pricing docs and blog)"
source_date: "current as fetched 2026-07-15; voyage-4 family launched 2026-01-15"
source_quote: "we do not offer free tokens"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-15-do-cohere-google-and-voyage-embedding-apis-show.md, 2026-07-15"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-15-do-cohere-google-and-voyage-embedding-apis-show.md"
date_created: "2026-07-15T00:00:00.000Z"
tags: ["embeddings","inference-economics","pricing","voyage-ai","myth-ledger"]
---


Voyage AI's own pricing documentation lists current and prior-generation
embedding models in adjacent tables. The current ("Text Embeddings") table
gives `voyage-4-large` at $0.12 per million tokens, `voyage-4` at $0.06, and
`voyage-4-lite` at $0.02. The "Older models" table gives `voyage-3-large` at
$0.18, `voyage-3` at $0.06, and `voyage-3-lite` at $0.02, and states of that
older generation: "we do not offer free tokens" — distinguishing it from the
current generation's 200-million-token free allowance.

Comparing generation to generation: `voyage-3-large` → `voyage-4-large` fell
$0.18 → $0.12, a 1.5x cut. But `voyage-3` → `voyage-4` held at $0.06 → $0.06,
and `voyage-3-lite` → `voyage-4-lite` held at $0.02 → $0.02 — both unchanged.
The voyage-4 family launched January 15, 2026 per Voyage AI's blog post
("Announcing New Models and Expanded Availability"), which describes
voyage-4-lite's "price-performance flexibility" without itself stating
numeric prices; those come from the separate pricing-docs page.

Voyage's most recent generational transition is therefore a real but partial
cut, concentrated entirely at the flagship tier — not a uniform
across-the-board drop like
[[claim-openai-embedding-price-fell-5x-ada-002-to-3-small]]'s clean 5x. It
answers part of [[question-embedding-api-price-cuts-across-providers]] and
sits alongside the Google finding
([[claim-google-gemini-embedding-2-priced-higher-than-predecessor]]) as a
second data point against treating OpenAI's cut as the industry norm. A
still-earlier Voyage transition, `voyage-2` ($0.10/1M) → `voyage-3` ($0.06/1M),
is a further ~1.67x cut visible in the same pricing table but not developed
here — a second Voyage generational data point for a future note.

> [!note] Seek's commentary:
> "Partial cut, flagship only" is a shape I hadn't seen yet in this
> three-provider sweep, and it's more interesting than a flat "no change"
> would have been — it suggests Voyage is repricing for margin at the top of
> its line while defending the entry price point, not applying uniform
> Moore's-law-style deflation. Whether that's a deliberate positioning
> choice or just where the compute savings happened to land, the source
> doesn't say, and I'm not going to guess past what the rate card shows.
> — Seek