Embedding Models

We support multiple embedding models. You can find the complete list along with API reference links at the end of the page.

What are Embeddings?

Embeddings from AI/ML API quantify the similarity between text strings. These embeddings are particularly useful for:

• Search: Rank search results by their relevance to a query.

• Clustering: Group similar text strings together.

• Recommendations: Suggest items based on related text strings.

• Anomaly Detection: Identify outliers that differ significantly from the norm.

• Diversity Measurement: Analyze the spread of similarities within a dataset.

• Classification: Categorize text strings by comparing them to labeled examples.

An embedding is a vector (list) of floating-point numbers, where the distance between vectors indicates their relatedness. Smaller distances indicate higher similarity, while larger distances suggest lower similarity.

Pricing

For more information on Embeddings pricing, visit our pricing page. Costs are calculated based on the number of tokens in the input.

Example: Generating Embeddings

curl https://api.aimlapi.com/v1/embeddings \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AIMLAPI_API_KEY" \
  -d '{
    "input": "Your text string goes here",
    "model": "text-embedding-3-small"
  }'

The response will include the embedding vector and additional metadata:

Response

{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "index": 0,
      "embedding": [
        -0.006929283495992422,
        -0.005336422007530928,
        // ...(omitted for spacing)
        -4.547132266452536e-05,
        -0.024047505110502243
      ]
    }
  ],
  "model": "text-embedding-3-small",
  "usage": {
    "prompt_tokens": 5,
    "total_tokens": 5
  }
}

By default, the length of the embedding vector is 1536 for text-embedding-3-small or 3072 for text-embedding-3-large. You can reduce the dimensions of the embedding using the dimensions parameter without losing its ability to represent concepts. More details on embedding dimensions can be found in the embedding use case section.

Example in Python

Here's how to use the embeddings API in Python:

import os
import json
import openai

# Initialize the API client
client = openai.OpenAI(
    base_url="https://api.aimlapi.com/v1",
    api_key=os.getenv("<YOUR_AIMLAPI_KEY"),
)

# Define the text for which to generate an embedding
text = "Your text string goes here"

# Request the embedding
response = client.embeddings.create(
    input=text,
    model="text-embedding-3-small"
)

# Extract the embedding from the response
embedding = response['data'][0]['embedding']

# Print the embedding
print(json.dumps(embedding, indent=2))

This Python example shows how to set up an API client, send text to the embeddings API, and handle the response to extract and print the embedding vector.

All Available Embedding Models

Model ID	Developer	Context	Model Card
text-embedding-3-small	Open AI	8000	-
text-embedding-3-large	Open AI	8000	Text-embedding-3-large
text-embedding-ada-002	Open AI	8000	Text-embedding-ada-002
togethercomputer/m2-bert-80M-32k-retrieval	Together AI	32000	M2-BERT-Retrieval-32k
BAAI/bge-base-en-v1.5	BAAI		BAAI-Bge-Base-1p5
togethercomputer/m2-bert-80M-2k-retrieval	Together AI		M2-BERT-Retrieval-2K
BAAI/bge-large-en-v1.5	BAAI		bge-large-en
togethercomputer/m2-bert-80M-8k-retrieval	Together AI	8000	M2-BERT-Retrieval-8k
voyage-large-2-instruct	Anthropic	16000	Voyage Large 2 Instruct
voyage-finance-2	Anthropic	32000	-
voyage-multilingual-2	Anthropic	32000	-
voyage-law-2	Anthropic	16000	-
voyage-code-2	Anthropic	16000	-
voyage-large-2	Anthropic	16000	-
voyage-2	Anthropic	4000	-
textembedding-gecko@003	Google	2000	Textembedding-gecko@003
textembedding-gecko-multilingual@001	Google	2000	Textembedding-gecko-multilingual@001
text-multilingual-embedding-002	Google	2000	-