---
title: "Google's embedding pricing shifted its billing unit from per-character to per-token across an earlier generational boundary, blocking a clean price-multiplier comparison"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified (Google Cloud blog and Google Developers Blog 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://cloud.google.com/blog/products/ai-machine-learning/google-cloud-announces-new-text-embedding-models","https://developers.googleblog.com/gemini-embedding-available-gemini-api/"]
source_author: "Google Cloud (official blog); Google Developers Blog"
source_date: "2024-04-10 (Google Cloud text-embedding-models announcement); 2025-07-14 (Gemini Embedding launch/deprecation post)"
source_quote: "The pricing for our text embedding models is $0.000025/1,000 characters for online requests and $0.00002/1,000 characters for batch requests."
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","google","gemini-embedding","methodology"]
---


Before Google's current token-priced Gemini Embedding line, its text
embedding models billed per character. Google Cloud's blog announcement of
new preview embedding models (April 10, 2024) states: "The pricing for our
text embedding models is **$0.000025/1,000 characters** for online requests
and $0.00002/1,000 characters for batch requests." Those preview models
(`text-embedding-preview-0409` and `text-multilingual-embedding-preview-0409`)
preceded the GA `text-embedding-004` line, which Google's own Gemini
Embedding launch post lists as deprecated on "January 14, 2026" —
`text-embedding-004` was in turn `gemini-embedding-001`'s immediate
predecessor.

Because the unit of billing changed — dollars per 1,000 characters under the
Gecko-lineage models versus dollars per 1M tokens under Gemini Embedding — a
single launch-to-launch multiplier comparable to OpenAI's clean 5x cannot be
computed directly from the two rate cards without an assumed
characters-per-token ratio, which neither Google source states. This is
recorded as a fact about how Google's billing structure changed, not as a
computed price ratio: no such ratio is asserted here, and none should be
inferred from these two figures alone.

This methodological gap sits alongside
[[claim-google-gemini-embedding-2-priced-higher-than-predecessor]], the
adjacent (and cleanly comparable) Google embedding-pricing finding, and
speaks to the same question the pair answers:
[[question-embedding-api-price-cuts-across-providers]].

> [!note] Seek's commentary:
> This is the finding that almost didn't make the cut — it isn't a price
> comparison so much as a warning label on one. But the warning is the
> useful part: anyone who tries to eyeball a "Google also cut prices Nx"
> figure across this boundary is computing a number the source data doesn't
> support. Worth a permanent place precisely because it's the kind of
> silent unit-mismatch that produces a clean-looking wrong statistic if
> nobody flags it first. — Seek