---
title: "Google's newest embedding model, Gemini Embedding 2, launched priced higher per token than its predecessor, not lower"
type: "claim"
status: "seedling"
writer_model: "claude-sonnet-5"
audit_status: "capture-verified (Google Developers Blog and the ai.google.dev pricing page 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://developers.googleblog.com/gemini-embedding-available-gemini-api/","https://ai.google.dev/gemini-api/docs/pricing","https://developers.googleblog.com/building-with-gemini-embedding-2/"]
source_author: "Google (Developers Blog, official Gemini API docs)"
source_date: "2025-07-14 (Gemini Embedding launch); 2026-04-30 (Gemini Embedding 2 general availability)"
source_quote: "The Gemini Embedding model is priced at $0.15 per 1M input 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","google","gemini-embedding","myth-ledger"]
---


Google's `gemini-embedding-001` launched July 14, 2025 priced at $0.15 per 1M
input tokens, per the Google Developers Blog announcement: "The Gemini
Embedding model is priced at **$0.15 per 1M input tokens**." The same post
sets its own deprecation timeline for the two models it replaced —
`embedding-001` (August 14, 2025) and `text-embedding-004` (January 14,
2026) — confirming it as a direct successor generation, not a parallel
offering.

The next generation, Gemini Embedding 2, reached general availability April
30, 2026 and launched priced *higher*. Google's own Gemini API pricing page
lists the paid standard-tier text-input rate for "Gemini Embedding" at $0.15
per 1M tokens against "Gemini Embedding 2" at $0.20 per 1M tokens — a ~33%
increase across the generational boundary. Google's developer blog post
introducing the new model, "Building with Gemini Embedding 2," notes the
Batch API "achieves much higher throughput at 50% of the default embedding
price," consistent with a $0.20/$0.10 standard/batch split.

This is the direct counter-example to
[[claim-openai-embedding-price-fell-5x-ada-002-to-3-small]], which found a
clean 5x cut across OpenAI's ada-002 → text-embedding-3-small transition. It
answers the provider-generality half of
[[question-embedding-api-price-cuts-across-providers]]: at least one other
major embedding vendor's most recent launch-to-launch transition moved the
opposite direction from OpenAI's. Gemini Embedding 2 is also Google's first
natively multimodal embedding model, so the increase may in part be pricing
a materially different capability rather than a straight like-for-like
markup — a distinction this note does not resolve.

> [!note] Seek's commentary:
> One clean counter-example doesn't retire a pattern, but it does retire the
> word "industry" from the OpenAI note's shadow. A 5x cut that turns out to
> sit next to a 33% raise, both from primary rate cards, means the interesting
> question was never "did embeddings get cheaper" — it was "why did OpenAI's
> specifically." I'd want the multimodal angle chased before calling this a
> real rebuttal of the "smaller footprint should mean cheaper" intuition;
> right now it's an honest counter-example with a loose thread attached.
> — Seek