Gemini Embedding 2 is Google's first multimodal embedding model, released 2026-05, mapping text and images into a unified vector space for semantic search and RAG. It supports output dimensions from 128 to 3,072, with recommended sizes of 768, 1536, or 3,072. The model enables cross-modal retrieval—embedding a text query to find relevant images or the reverse—making it suited for multimodal search, recommendation, and document understanding.
Channels
Every channel serves the same model, but they differ in price and stability. On each API key in the Platform, you pick a primary and a backup channel per vendor, and each billing line shows which channel served the request.
Context pricing
This model’s price depends on how much input context a request carries. The headline price is the cheapest band, and a longer prompt moves the whole request into a higher one.
Capabilities
Quick start
POST /v1/embeddingsimport os from openai import OpenAI client = OpenAI( base_url="https://api.mirapi.ai/v1", # changed api_key=os.environ["MIRAPI_API_KEY"], # changed ) resp = client.embeddings.create( model="gemini-embedding-2", input="The quick brown fox", ) vector = resp.data[0].embedding # len(vector) — pass dimensions= to shorten it
import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://api.mirapi.ai/v1", // changed apiKey: process.env.MIRAPI_API_KEY, // changed }); const resp = await client.embeddings.create({ model: "gemini-embedding-2", input: "The quick brown fox", }); const vector = resp.data[0].embedding;
curl https://api.mirapi.ai/v1/embeddings \ -H "Authorization: Bearer $MIRAPI_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gemini-embedding-2", "input": "The quick brown fox" }'
Links
1M free tokens to start, no card required. It’s OpenAI-compatible, so you only change the base URL.