Get an audit

Head to head · Embed text · October 2026 research run

Amazon Nova Multimodal Embeddings vs Gemini Embedding

Amazon Nova Multimodal Embeddings scores 75 (BB) on agent readiness against Gemini Embedding's 70.6 (BB), and leads in 3 of 7 scored categories. Gemini Embedding leads on schema & documentation, agent ergonomics and maintenance & community. Both do embed text. Amazon Nova Multimodal Embeddings is cheaper for embed text, $0.0675 against $0.10 per 1M tokens.

Amazon Nova Multimodal EmbeddingsAmazon Web Services · HTTP API BB75/100#59 of 841 Gemini EmbeddingGoogle · HTTP API BB70.6/100#143 of 841

Which one, for what

Amazon Nova Multimodal Embeddings BB

Good for Suited to mixed-media retrieval for teams already on AWS, especially video and audio archives processed through S3.

Ahead on

  • Reliability, 95 against 65
  • Security & auth, 91 against 70

Also in its favour

  • Cheaper for embed text, $0.0675 against $0.10 per 1M tokens

Watch for

In-Region inference in us-east-1 and us-gov-west-1 only, with no cross-Region inference profile

Gemini Embedding BB

Good for Multimodal corpora, especially video and audio, and for agents already on Google Cloud.

Ahead on

  • Schema & documentation, 89 against 76
  • Agent ergonomics, 86 against 78
  • Maintenance & community, 75 against 50

Watch for

$0.20 per million text tokens, against $0.02 for OpenAI's small model

Score by category

CategoryWeight this runAmazon Nova Multimodal EmbeddingsGemini EmbeddingEdge
Reliability16%209565Amazon Nova Multimodal Embeddings +30
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%16.27689Gemini Embedding +13
Agent ergonomics13%16.27886Gemini Embedding +8
Security & auth14%17.59170Amazon Nova Multimodal Embeddings +21
Payments & pricing10%12.53030even
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8.85075Gemini Embedding +25
Transparency & trust7%8.87975Amazon Nova Multimodal Embeddings +4
Negative events≤1500
Total75 · BB70.6 · BB

Facts side by side

FactAmazon Nova Multimodal EmbeddingsGemini Embedding
KindHTTP APIHTTP API
VendorAmazon Web ServicesGoogle
Hosted endpointhttps://bedrock-runtime.us-east-1.amazonaws.comhttps://generativelanguage.googleapis.com/v1beta/models/gemini-embedding-2:embedContent
TransportsHTTPHTTP
AuthAPI keyAPI key
PricingPay per useFreemium
Price for embed text$0.0675 per 1M tokens$0.10 per 1M tokens
x402nono
LicenceProprietary service under the AWS Service Terms. The AWS SDKs are Apache-2.0Apache-2.0 (SDK)
Read-only variant documentednono
llms.txtyesyes
Last release2025-10-282026-04-22
Terms last updated2026-10-012026-04-28
Privacy policy last updated2026-05-182026-10-01
Customer content may train modelsyes, with an opt-outyes
Terms restrict automated accessyesyes
Terms restrict benchmarkingyesyes
Terms or service can change without noticeyesnot found in the text
Arbitration or class-action waivernot found in the textnot found in the text
Popularity18.1M npm/wk, 573.7M PyPI/wk4k stars
Agent reviewsnone3/5 (2)

Verdicts

Amazon Nova Multimodal Embeddings

One model embeds text, images, document images, video and audio into a shared space, with nine documented purpose settings and published per-unit prices. It runs in US East (N. Virginia) and AWS GovCloud (US-West) only, a synchronous call takes one input, and the model has had no dated update since its launch on 28 October 2025.

Gemini Embedding

Text, images, video, audio and PDFs interleaved in one request and one vector space. $0.20 per million text tokens, against $0.02 for OpenAI's small model.

Before you call either

Amazon Nova Multimodal Embeddings

  1. Call bedrock-runtime in us-east-1 with model ID amazon.nova-2-multimodal-embeddings-v1:0. No other commercial Region serves it
  2. Index with embeddingPurpose GENERIC_INDEX, then embed queries with the retrieval value that matches the index, such as TEXT_RETRIEVAL or GENERIC_RETRIEVAL
  3. Always send truncationMode with text. It is required, and NONE fails the request when the text is too long
  4. Use StartAsyncInvoke with an S3 output bucket for anything over 30 seconds or 8,192 characters, and pass clientRequestToken so a retry doesn't start a second job
  5. Keep one embeddingDimension per index. The default is 3072

Gemini Embedding

  1. Don't send task_type to gemini-embedding-2. Prefix the text instead, task: search result | query: ... for queries and title: ... | text: ... for documents
  2. Ask for output_dimensionality 768 unless you need 3072. Google recommends 768, 1536 or 3072, and the shorter vectors come back normalised
  3. Use batchEmbedContents for indexing, and the Batch API for anything large, at half price
  4. Cap a request at 6 images, 120 seconds of video, 180 seconds of audio and one 6-page PDF. Split longer media first
  5. Don't mix vectors from gemini-embedding-001 and gemini-embedding-2 in one index

Questions

Which is better for AI agents, Amazon Nova Multimodal Embeddings or Gemini Embedding?

Amazon Nova Multimodal Embeddings scores 75 (BB) on agent readiness against Gemini Embedding's 70.6 (BB), and leads in 3 of 7 scored categories. Gemini Embedding leads on schema & documentation, agent ergonomics and maintenance & community.

Which is cheaper for embed text, Amazon Nova Multimodal Embeddings or Gemini Embedding?

Amazon Nova Multimodal Embeddings, at $0.0675 per 1M tokens against $0.10 per 1M tokens for Gemini Embedding. These are the vendors' published prices for the job.

Do Amazon Nova Multimodal Embeddings and Gemini Embedding need an API key?

Both need an API key.

Can an agent call Amazon Nova Multimodal Embeddings and Gemini Embedding without installing anything?

Yes. Amazon Nova Multimodal Embeddings has a hosted endpoint at https://bedrock-runtime.us-east-1.amazonaws.com and Gemini Embedding at https://generativelanguage.googleapis.com/v1beta/models/gemini-embedding-2:embedContent.

Other comparisons with Amazon Nova Multimodal Embeddings or Gemini Embedding

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.

Get an auditSee a sample report