Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Balance
Subscription
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksEarn crypto

Private AI

Explore Embedding Models

Compare AI embedding and reranking models for search and RAG by price, dimensions, token limit, image input, batch and Zero Data Retention support.

Models

How embedding models work

  • Embeddings turn text, and on some models images, into vectors for search, RAG and clustering. Call POST /api/v1/embeddings.
  • Reranking scores documents against a query. Call POST /api/v1/rerank.
  • Cohere Embed 5 and Nemotron 3.5 also need input_type.
  • API only: they cannot be used in chat. Prices are per 1M input tokens.
Embeddings API docs Rerank API docs List models as JSON Batch APIGet an API key

Example request body

{
  "model": "text-embedding-3-small",
  "input": [
    "Where is my order?",
    "Track a delivery"
  ]
}

Embedding models (17)

  • Cohere Embed 5 Fast

    Cohere's low-latency embedding model for interactive search and high-volume retrieval. Embed 5 Fast and Pro share a vector space: index with Pro and query with Fast using the same dimensions. Supports text, images, mixed text/image items, over 100 languages, and a 128,000-token context. Choose 256, 512, 768, 1024, 1536, or 2048 dimensions (default), with float or base64 output. Set input_type to search_query, search_document, classification, clustering, or image. Send up to 96 items per request. Images use PNG, JPEG, or WebP URLs or data URLs, with up to 16 images, 8 MiB per image and 16 MiB combined; images are resized to fit 1,120 pixels. Text and image tokens are billed separately. Image requests require account or team balance; x402 is not supported.
    Price
    $0.15/M
    Image tokens
    $0.50/M
    Max tokens
    128K
    Dimensions
    256, 512, 768, 1,024, 1,536, 2,048
    Batch API
    No
    ZDR
    —

    Added Nov 3, 2026

  • Cohere Embed 5 Pro

    Cohere's embedding model for high-quality document indexing and retrieval. Embed 5 Fast and Pro share a vector space: index with Pro and query with Fast using the same dimensions. Supports text, images, mixed text/image items, over 100 languages, and a 128,000-token context. Choose 256, 512, 768, 1024, 1536, or 2048 dimensions (default), with float or base64 output. Set input_type to search_query, search_document, classification, clustering, or image. Send up to 96 items per request. Images use PNG, JPEG, or WebP URLs or data URLs, with up to 16 images, 8 MiB per image and 16 MiB combined; images are resized to fit 1,120 pixels. Text and image tokens are billed separately. Image requests require account or team balance; x402 is not supported.
    Price
    $0.15/M
    Image tokens
    $0.50/M
    Max tokens
    128K
    Dimensions
    256, 512, 768, 1,024, 1,536, 2,048
    Batch API
    No
    ZDR
    —

    Added Nov 3, 2026

  • Nemotron 3.5 Embed 1B

    NVIDIA multilingual retrieval model that converts text queries and text or image documents into L2-normalized vectors with 2,048 native dimensions. Supports reduced dimensions from 32 to 2,048 and a 32,768-token context. Set input_type to query or document and send text without query: or passage: prefixes. Document items may contain one image with optional text, with up to 16 images per request. All tokens in an image-containing item use the image rate; standalone text items use the text rate. Image requests require account or team balance; x402 is not supported.
    Price
    $0.015/M
    Image items
    $0.035/M
    Max tokens
    32.8K
    Dimensions
    32–2,048
    Batch API
    No
    ZDR
    —

    Added Oct 21, 2026

  • Nemotron 3.5 Embed 8B

    NVIDIA multilingual retrieval model that converts text queries and text or image documents into L2-normalized vectors with 4,096 native dimensions. Supports reduced dimensions from 32 to 4,096 and a 32,768-token context. Set input_type to query or document and send text without query: or passage: prefixes. Document items may contain one image with optional text, with up to 16 images per request. All tokens in an image-containing item use the image rate; standalone text items use the text rate. Image requests require account or team balance; x402 is not supported.
    Price
    $0.035/M
    Image items
    $0.07/M
    Max tokens
    32.8K
    Dimensions
    32–4,096
    Batch API
    No
    ZDR
    —

    Added Oct 21, 2026

  • Gemini Embedding 2

    Google Gemini Embedding 2 model with text/image input and dimension reduction support
    Price
    $0.20/M
    Images
    Yes
    Max tokens
    8.2K
    Dimensions
    128–3,072
    Batch API
    No
    ZDR
    —

    Added Mar 10, 2026

  • Gemini Embedding 2 Preview

    Google Gemini Embedding 2 preview model with text/image input and dimension reduction support
    Price
    $0.20/M
    Images
    Yes
    Max tokens
    8.2K
    Dimensions
    128–3,072
    Batch API
    No
    ZDR
    —

    Added Mar 10, 2026

  • BGE Base English v1.5

    BAAI English embedding model (base)
    Price
    $0.005/M
    Images
    No
    Max tokens
    512
    Dimensions
    768
    Batch API
    No
    ZDR

    Added Sep 16, 2025

  • BGE Large English v1.5

    BAAI English embedding model
    Price
    $0.01/M
    Images
    No
    Max tokens
    512
    Dimensions
    1,024
    Batch API
    No
    ZDR

    Added Aug 6, 2025

  • BGE M3

    BAAI multilingual embedding model
    Price
    $0.01/M
    Images
    No
    Max tokens
    8.2K
    Dimensions
    1,024
    Batch API
    No
    ZDR

    Added Aug 6, 2025

  • Gemini Embedding

    Alias for Gemini Embedding (maps to gemini-embedding-001)
    Price
    $0.15/M
    Images
    No
    Max tokens
    8.2K
    Dimensions
    3,072
    Batch API
    No
    ZDR
    —

    Added Aug 6, 2025

  • Gemini Embedding 001

    Google Gemini embedding model - High-performance semantic embeddings
    Price
    $0.15/M
    Images
    No
    Max tokens
    8.2K
    Dimensions
    3,072
    Batch API
    No
    ZDR
    —

    Added Aug 6, 2025

  • Qwen3 Embedding 0.6B

    Qwen3 embedding model with dimension reduction support
    Price
    $0.01/M
    Images
    No
    Max tokens
    8.2K
    Dimensions
    Up to 1,024
    Batch API
    No
    ZDR

    Added Aug 6, 2025

  • Qwen3 Embedding 4B

    Qwen3 4B parameter embedding model with dimension reduction support
    Price
    $0.02/M
    Images
    No
    Max tokens
    32.8K
    Dimensions
    32–2,560
    Batch API
    No
    ZDR

    Added Aug 6, 2025

  • Qwen3 Embedding 8B

    Qwen3 8B parameter embedding model with dimension reduction support
    Price
    $0.01/M
    Images
    No
    Max tokens
    32.8K
    Dimensions
    Up to 4,096
    Batch API
    No
    ZDR

    Added Aug 6, 2025

  • Text Embedding 3 Large

    Highest performance OpenAI embedding model with dimension reduction support
    Price
    $0.13/M
    Images
    No
    Max tokens
    8.2K
    Dimensions
    Up to 3,072
    Batch API
    Yes
    ZDR
    —

    Added Aug 6, 2025

  • Text Embedding 3 Small

    Most cost-effective OpenAI embedding model with dimension reduction support
    Price
    $0.02/M
    Images
    No
    Max tokens
    8.2K
    Dimensions
    Up to 1,536
    Batch API
    Yes
    ZDR
    —

    Added Aug 6, 2025

  • Text Embedding Ada 002

    Legacy OpenAI embedding model
    Price
    $0.10/M
    Images
    No
    Max tokens
    8.2K
    Dimensions
    1,536
    Batch API
    Yes
    ZDR
    —

    Added Aug 6, 2025

Reranking models (1)

  • Nemotron 3.5 Rerank 8B

    NVIDIA multilingual reranker that scores text documents against a text query. Returns raw relevance logits, with higher scores indicating greater relevance, sorted from highest to lowest while preserving original document indexes. Supports up to 1,024 documents and a 32,768-token context per query-document pair. Use top_n to return the best matches. This API currently accepts text only.
    Price
    $0.05/M
    Images
    No
    Max tokens
    32.8K
    Documents
    1,024
    Batch API
    No
    ZDR
    —

    Added Oct 21, 2026