Models served via Google Vertex AI

5 models from 1 makers that you can call through Google Vertex AI — each one tested by us with fixed prompts, outputs kept, costs measured.

Makers: google

5 models 🧪 all independently tested

gemma-3-4b-it

131K

gemini-2.0-flash-preview-image-generation

33K

imagen-3.0-generate-002

480

Imagen 3.0 Image Editing

N/A

Imagen (latest stable)

N/A

About Google Vertex AI

first party

Vertex AI is Google's enterprise cloud channel for the same models available through the direct Gemini API — the difference is billing and controls, not the models themselves. Vertex AI usage bills through Google Cloud rather than a separate personal API key, and adds enterprise SLAs, IAM, and VPC controls.

If you're evaluating Google's models for an individual project, the direct Gemini API (see that provider page) is usually the simpler starting point; Vertex AI is aimed at teams already standardized on Google Cloud.

  • Google's enterprise cloud channel for Gemini and other first-party models
  • Bills through Google Cloud billing rather than a separate personal API key
  • Adds enterprise SLAs, IAM, and VPC controls the direct Gemini API doesn't offer

Affiliate program: Same Google Cloud affiliate program as the direct Gemini API — 5% via CJ Affiliate, Google-Cloud-wide, not specific to any one product. Program details

Google Vertex AI pricing, by model

This is what Google Vertex AI itself charges for each model — not the cheapest price across every provider (that's what the cards above show). Prompt tokens range $0.04–$0.04 per 1M across 1 priced models.

ModelMakerPrice via Google Vertex AIContext
gemini-2.0-flash-preview-image-generationgoogle33K
gemma-3-4b-itgoogle$0.04 / $0.08 per 1M33K
imagen-3.0-generate-002google480
Imagen 3.0 Image Editinggoogle
Imagen (latest stable)google

Latency: Per-provider latency isn't independently measured yet — the probe rig currently tests one route per model, not one per provider (see PROBE_RIG.md ledger #10).

FAQ

Why are there two Google provider pages?
“Google” is both a model maker and, separately, an api_provider — the direct Gemini API and Vertex AI are two different commercial channels for the same underlying models, so they get two provider pages here rather than one page pretending to be a single offering.