---
title: "Embedding and reranking models"
description: "Compare embedding models and rerankers for RAG, from Voyage AI, Cohere, and OpenAI to open models like BGE and Qwen3, by license, modality, and hosting."
type: "platform-category"
canonical_url: "https://theaiengineering.com/platforms/embedding-reranking"
markdown_url: "https://theaiengineering.com/platforms/embedding-reranking.md"
topic: "RAG and vector search"
author: "The AI Engineering Editors"
publisher: "The AI Engineering"
citation: "The AI Engineering Editors. \"Embedding and reranking models.\" The AI Engineering. https://theaiengineering.com/platforms/embedding-reranking"
---

# Embedding and reranking models

An embedding model converts text, images, or other content into vectors so that items with similar meaning sit close together, and a reranker scores a query against each candidate passage to reorder the results a retriever returns. Together they decide which context a RAG system passes to the language model.

Options include hosted APIs from Voyage AI, Cohere, OpenAI, Google, Amazon, and Mistral AI; open-weight families such as BGE, Qwen3 Embedding, Nomic Embed, Snowflake Arctic Embed, and IBM Granite Embedding; and libraries and servers, such as Sentence Transformers and Text Embeddings Inference, that run open models on your own hardware.

Public leaderboards such as MTEB are a useful starting point, but scores on general benchmarks often differ from results on a team's own documents. Licenses vary as well: several open-weight embedding models and rerankers are released for non-commercial use only.

## Embedding and reranking models compared

| Platform | Open source | Deployment | Best for |
| --- | --- | --- | --- |
| [Amazon Nova and Titan embeddings](https://aws.amazon.com/nova/models/) | No | API, Managed cloud | Retrieval applications built on Amazon Bedrock |
| [BGE (BAAI General Embedding)](https://bge-model.com) | Yes (MIT) | Library, Self-hosted | Self-hosted multilingual hybrid retrieval |
| [Cohere Embed and Rerank](https://cohere.com/embed) | No | API, Managed cloud, Self-hosted | Multilingual enterprise retrieval and reranking |
| [Contextual AI Reranker](https://docs.contextual.ai/how-to-guides/rerank) | No | API | Reranking with natural-language ranking instructions |
| [EmbeddingGemma](https://ai.google.dev/gemma/docs/embeddinggemma) | Yes (Gemma Terms of Use) | Local, Self-hosted | On-device and offline embeddings |
| [FastEmbed](https://qdrant.github.io/fastembed/) | Yes (Apache-2.0) | Library | Local embedding generation for Qdrant pipelines |
| [Gemini Embedding](https://ai.google.dev/gemini-api/docs/embeddings) | No | API, Managed cloud | Multimodal retrieval on Google's stack |
| [IBM Granite Embedding](https://github.com/ibm-granite/granite-embedding-models) | Yes (Apache-2.0) | Self-hosted, Local | Permissively licensed embeddings for enterprise use |
| [Infinity (embedding server)](https://github.com/michaelfeil/infinity) | Yes (MIT) | Self-hosted | Self-hosting several embedding and reranking models together |
| [Jina AI embeddings and rerankers](https://jina.ai/embeddings/) | No | API, Managed cloud | Multilingual and multimodal embeddings with a hosted API |
| [Mistral embeddings](https://docs.mistral.ai/inference/embeddings) | No | API | Code retrieval and teams using Mistral models |
| [Mixedbread](https://www.mixedbread.com) | No | Managed cloud, API | Open rerankers or managed multimodal search |
| [Nomic Embed](https://huggingface.co/nomic-ai) | Yes (Apache-2.0) | Self-hosted, Local | Lightweight open embeddings for local pipelines |
| [NVIDIA Nemotron Retriever](https://huggingface.co/collections/nvidia/nemotron-retriever) | Yes (NVIDIA Open Model License) | Self-hosted, API | GPU-accelerated retrieval on NVIDIA infrastructure |
| [OpenAI embeddings](https://developers.openai.com/api/docs/guides/embeddings) | No | API | Teams already building on the OpenAI API |
| [Perplexity pplx-embed](https://docs.perplexity.ai/docs/embeddings/quickstart) | Yes (MIT) | API, Self-hosted | Open embeddings for web-scale retrieval |
| [PyLate](https://lightonai.github.io/pylate/) | Yes (MIT) | Library | Training and serving ColBERT-style models |
| [Qwen3 Embedding and Reranker](https://github.com/QwenLM/Qwen3-Embedding) | Yes (Apache-2.0) | Self-hosted, Local | Open-weight multilingual embeddings at several sizes |
| [Sentence Transformers](https://www.sbert.net) | Yes (Apache-2.0) | Library | Training and running open embedding models in Python |
| [Snowflake Arctic Embed](https://github.com/Snowflake-Labs/arctic-embed) | Yes (Apache-2.0) | Self-hosted, Local | Open retrieval embeddings and Snowflake users |
| [Text Embeddings Inference](https://huggingface.co/docs/text-embeddings-inference) | Yes (Apache-2.0) | Self-hosted | Serving open embedding models in production |
| [Voyage AI](https://www.voyageai.com) | No | API, Managed cloud | Retrieval-tuned embeddings, especially alongside MongoDB |
| [ZeroEntropy](https://www.zeroentropy.dev) | No | API, Managed cloud | Adding a reranking stage to existing retrieval |

## Profiles

### Amazon Nova and Titan embeddings

Amazon Bedrock offers Amazon Nova Multimodal Embeddings, a single model for text, documents, images, video, and audio launched in October 2025, alongside Titan Text Embeddings V2 and Titan Multimodal Embeddings for text and image retrieval.

- Company: Amazon Web Services
- Website: https://aws.amazon.com/nova/models/
- Open source: No
- Deployment: API, Managed cloud
- Best for: Retrieval applications built on Amazon Bedrock
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#amazon-nova-embeddings

### BGE (BAAI General Embedding)

BGE is a family of open embedding and reranking models from BAAI, including bge-m3 for dense, sparse, and multi-vector retrieval in more than 100 languages and the bge-reranker-v2 series, distributed with the FlagEmbedding library.

- Company: Beijing Academy of Artificial Intelligence (BAAI)
- Website: https://bge-model.com
- GitHub: https://github.com/FlagOpen/FlagEmbedding
- Open source: Yes (MIT)
- Deployment: Library, Self-hosted
- Languages: Python
- Best for: Self-hosted multilingual hybrid retrieval
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#bge

### Cohere Embed and Rerank

Cohere's Embed v4 produces multilingual text and image embeddings with Matryoshka dimensions and long context, and Rerank 4, in Pro and Fast variants, reorders search results; both run on Cohere's API, major clouds, and private deployments.

- Company: Cohere
- Website: https://cohere.com/embed
- Open source: No
- Deployment: API, Managed cloud, Self-hosted
- Languages: Python, TypeScript, Go, Java
- Best for: Multilingual enterprise retrieval and reranking
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#cohere-embed-rerank

### Contextual AI Reranker

Contextual AI's Reranker v2 is an instruction-following, multilingual reranker family in 1B, 2B, and 6B sizes, available through the Contextual AI API and as Hugging Face weights under the non-commercial CC-BY-NC-SA-4.0 license.

- Company: Contextual AI
- Website: https://docs.contextual.ai/how-to-guides/rerank
- Open source: No
- Deployment: API
- Best for: Reranking with natural-language ranking instructions
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#contextual-ai-reranker

### EmbeddingGemma

EmbeddingGemma is a 308M-parameter open-weight embedding model from Google DeepMind, built on Gemma 3 and designed to run on-device, with Matryoshka output sizes and training data covering more than 100 languages.

- Company: Google DeepMind
- Website: https://ai.google.dev/gemma/docs/embeddinggemma
- Open source: Yes (Gemma Terms of Use)
- Deployment: Local, Self-hosted
- Best for: On-device and offline embeddings
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#embeddinggemma

### FastEmbed

FastEmbed is Qdrant's lightweight Python library for generating dense, sparse (SPLADE and BM25), and late-interaction ColBERT embeddings with ONNX Runtime instead of PyTorch, on CPUs or GPUs.

- Company: Qdrant
- Website: https://qdrant.github.io/fastembed/
- GitHub: https://github.com/qdrant/fastembed
- Open source: Yes (Apache-2.0)
- Deployment: Library
- Languages: Python
- Best for: Local embedding generation for Qdrant pipelines
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#fastembed

### Gemini Embedding

Google's Gemini Embedding models, gemini-embedding-001 for text and the natively multimodal Gemini Embedding 2 for text, images, video, audio, and PDFs, are available through the Gemini API and Google Cloud with Matryoshka output dimensions.

- Company: Google
- Website: https://ai.google.dev/gemini-api/docs/embeddings
- Open source: No
- Deployment: API, Managed cloud
- Languages: Python, TypeScript
- Best for: Multimodal retrieval on Google's stack
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#gemini-embedding

### IBM Granite Embedding

Granite Embedding is IBM's family of Apache-2.0 embedding and reranker models, with English and multilingual versions and R2 models based on the ModernBERT architecture, aimed at enterprise retrieval.

- Company: IBM
- Website: https://github.com/ibm-granite/granite-embedding-models
- GitHub: https://github.com/ibm-granite/granite-embedding-models
- Open source: Yes (Apache-2.0)
- Deployment: Self-hosted, Local
- Best for: Permissively licensed embeddings for enterprise use
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#granite-embedding

### Infinity (embedding server)

Infinity is an MIT-licensed server for embedding, reranking, and classification models behind a REST API, with a native Rust and LibTorch v3 server released in October 2026 alongside the earlier Python server and its ONNX and TensorRT options.

- Website: https://github.com/michaelfeil/infinity
- GitHub: https://github.com/michaelfeil/infinity
- Open source: Yes (MIT)
- Deployment: Self-hosted
- Languages: Rust, Python
- Best for: Self-hosting several embedding and reranking models together
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#infinity-embedding-server

### Jina AI embeddings and rerankers

Jina AI offers multilingual embedding models such as jina-embeddings-v5-text, the multimodal v5-omni and v4 models, and the jina-reranker-v3 and m0 rerankers through its API, with weights on Hugging Face mostly under non-commercial licenses.

Elastic completed its acquisition of Jina AI in October 2025; Jina models are also served through the Elastic Inference Service.

- Company: Elastic
- Website: https://jina.ai/embeddings/
- Open source: No
- Deployment: API, Managed cloud
- Best for: Multilingual and multimodal embeddings with a hosted API
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#jina-embeddings

### Mistral embeddings

Mistral AI's embeddings endpoint serves two models, mistral-embed for general text and codestral-embed for code retrieval, through the same API and SDKs as Mistral's generative and OCR models.

- Company: Mistral AI
- Website: https://docs.mistral.ai/inference/embeddings
- Open source: No
- Deployment: API
- Best for: Code retrieval and teams using Mistral models
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#mistral-embed

### Mixedbread

Mixedbread publishes Apache-2.0 models such as mxbai-embed-large-v1 and the mxbai-rerank-v2 rerankers on Hugging Face, and its current product, Mixedbread Search, wraps its own embedding and listwise reranking models in a managed retrieval API.

- Website: https://www.mixedbread.com
- Open source: No
- Deployment: Managed cloud, API
- Best for: Open rerankers or managed multimodal search
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#mixedbread

### Nomic Embed

Nomic Embed is a family of Apache-2.0 embedding models, including nomic-embed-text-v1.5 with Matryoshka dimensions, the multilingual mixture-of-experts v2, and nomic-embed-code; Nomic itself now builds AI software for architecture, engineering, and construction firms.

- Company: Nomic AI
- Website: https://huggingface.co/nomic-ai
- Open source: Yes (Apache-2.0)
- Deployment: Self-hosted, Local
- Best for: Lightweight open embeddings for local pipelines
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#nomic-embed

### NVIDIA Nemotron Retriever

Nemotron Retriever is NVIDIA's collection of open models for embedding, reranking, and document parsing, such as llama-nemotron-embed, packaged as NIM microservices within the NeMo Retriever stack for GPU-accelerated retrieval pipelines.

Formerly known as Nemotron RAG; NVIDIA also renamed individual NeMo Retriever models to Nemotron names.

- Company: NVIDIA
- Website: https://huggingface.co/collections/nvidia/nemotron-retriever
- Open source: Yes (NVIDIA Open Model License)
- Deployment: Self-hosted, API
- Best for: GPU-accelerated retrieval on NVIDIA infrastructure
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#nemotron-retriever

### OpenAI embeddings

OpenAI's embeddings API serves the text-embedding-3-small and text-embedding-3-large models, which return text embeddings whose length can be reduced with a dimensions parameter to trade some accuracy for smaller indexes.

- Company: OpenAI
- Website: https://developers.openai.com/api/docs/guides/embeddings
- Open source: No
- Deployment: API
- Languages: Python, TypeScript
- Best for: Teams already building on the OpenAI API
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#openai-embeddings

### Perplexity pplx-embed

pplx-embed is Perplexity's family of MIT-licensed embedding models, with v1 text models at 0.6B and 4B in standard and contextual variants served through Perplexity's API, and v2 late-interaction models for text, images, and PDFs.

- Company: Perplexity
- Website: https://docs.perplexity.ai/docs/embeddings/quickstart
- Open source: Yes (MIT)
- Deployment: API, Self-hosted
- Best for: Open embeddings for web-scale retrieval
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#pplx-embed

### PyLate

PyLate is LightOn's MIT-licensed library for training and running late-interaction retrieval models such as ColBERT, built on Sentence Transformers, with indexing and retrieval utilities for multi-vector search.

- Company: LightOn
- Website: https://lightonai.github.io/pylate/
- GitHub: https://github.com/lightonai/pylate
- Open source: Yes (MIT)
- Deployment: Library
- Languages: Python
- Best for: Training and serving ColBERT-style models
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#pylate

### Qwen3 Embedding and Reranker

Qwen3 Embedding and Qwen3 Reranker are Apache-2.0 open-weight models from Alibaba's Qwen team in 0.6B, 4B, and 8B sizes, joined by Qwen3-VL-Embedding and Qwen3-VL-Reranker for text, images, document images, and video.

- Company: Alibaba (Qwen team)
- Website: https://github.com/QwenLM/Qwen3-Embedding
- GitHub: https://github.com/QwenLM/Qwen3-Embedding
- Open source: Yes (Apache-2.0)
- Deployment: Self-hosted, Local
- Best for: Open-weight multilingual embeddings at several sizes
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#qwen3-embedding

### Sentence Transformers

Sentence Transformers is a Python library for computing, training, and fine-tuning embedding models, cross-encoder rerankers, and sparse encoders, originally from TU Darmstadt's UKP Lab and now maintained by Hugging Face.

- Company: Hugging Face
- Website: https://www.sbert.net
- GitHub: https://github.com/huggingface/sentence-transformers
- Open source: Yes (Apache-2.0)
- Deployment: Library
- Languages: Python
- Best for: Training and running open embedding models in Python
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#sentence-transformers

### Snowflake Arctic Embed

Arctic Embed is a family of Apache-2.0 text embedding models from Snowflake, including the multilingual snowflake-arctic-embed-l-v2.0 and m-v2.0 models, trained for retrieval and also callable inside Snowflake Cortex.

- Company: Snowflake
- Website: https://github.com/Snowflake-Labs/arctic-embed
- GitHub: https://github.com/Snowflake-Labs/arctic-embed
- Open source: Yes (Apache-2.0)
- Deployment: Self-hosted, Local
- Best for: Open retrieval embeddings and Snowflake users
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#arctic-embed

### Text Embeddings Inference

Text Embeddings Inference is Hugging Face's open-source Rust server for deploying embedding, reranker, and sequence classification models, with token-based dynamic batching, optimized kernels, and Docker images for CPUs and NVIDIA GPUs.

- Company: Hugging Face
- Website: https://huggingface.co/docs/text-embeddings-inference
- GitHub: https://github.com/huggingface/text-embeddings-inference
- Open source: Yes (Apache-2.0)
- Deployment: Self-hosted
- Languages: Rust
- Best for: Serving open embedding models in production
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#text-embeddings-inference

### Voyage AI

Voyage AI builds embedding and reranking models for retrieval, including the voyage-4 family with an open-weight voyage-4-nano, domain models for code, finance, and law, multimodal embeddings, and the rerank-2.5 and rerank-3 rerankers.

MongoDB acquired Voyage AI in 2025; the models are offered through the Voyage API and inside MongoDB Atlas.

- Company: MongoDB
- Website: https://www.voyageai.com
- Open source: No
- Deployment: API, Managed cloud
- Languages: Python, TypeScript
- Best for: Retrieval-tuned embeddings, especially alongside MongoDB
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#voyage-ai

### ZeroEntropy

ZeroEntropy trains retrieval models, including the zerank-2 reranker and the zembed-1 embedding model, offered through its API and cloud marketplaces, with Apache-2.0 weights for zerank-2, zerank-1-small, and zembed-1 on Hugging Face.

- Website: https://www.zeroentropy.dev
- Open source: No
- Deployment: API, Managed cloud
- Best for: Adding a reranking stage to existing retrieval
- Permalink: https://theaiengineering.com/platforms/embedding-reranking#zeroentropy

## How to choose

1. **Hosting**: A hosted API, open weights you run yourself, or both, which affects latency, data residency, and cost at scale.
2. **License**: Permissive licenses such as Apache-2.0 or MIT, custom model licenses, or non-commercial terms such as CC-BY-NC that require a separate agreement for production use.
3. **Modality and languages**: Text-only models versus multimodal models that embed images, PDFs, audio, or video, and the languages each model was trained to retrieve.
4. **Vector size and context**: Output dimensions, Matryoshka truncation, quantized output types, and maximum input length determine storage cost and chunking strategy.
5. **Retrieval style**: Single-vector dense embeddings, sparse lexical vectors, or late-interaction multi-vector models such as ColBERT, each of which needs different index support.
6. **Reranker design**: Cross-encoder, listwise, and instruction-following rerankers trade added latency for relevance, and some accept natural-language ranking instructions.

## Frequently asked questions

### What is an embedding model?

An embedding model is a neural network that maps text, images, or other inputs to fixed-length vectors, placing inputs with similar meaning close together in vector space. Retrieval systems embed documents once, store the vectors in a vector database, and embed each query at search time to find the nearest documents. The same vectors also support clustering, deduplication, and classification.

### What does a reranker do in RAG?

A reranker takes the query and a short list of candidate passages from the first retrieval step, often the top 50 to 100, and scores each pair directly for relevance. The highest-scoring passages are then passed to the language model. Because the reranker reads query and passage together, it catches relevance that vector similarity misses, at the cost of extra latency per query.

### What is the difference between a reranker and an embedding model?

An embedding model encodes queries and documents separately into vectors, so documents can be indexed in advance and searched quickly across millions of items. A reranker, usually a cross-encoder, processes the query and one document together and outputs a relevance score, which is more precise but too slow for a whole corpus. Most RAG pipelines use embeddings to retrieve candidates and a reranker to order them.

### What are the best open source embedding models for RAG?

Widely used open-weight families include BGE from BAAI, Qwen3 Embedding from Alibaba, Nomic Embed, Snowflake Arctic Embed, IBM Granite Embedding, Google's EmbeddingGemma, and Perplexity's pplx-embed. Check each license before production use, since some open-weight models, including several from Jina AI and Contextual AI, are non-commercial. Then compare candidates on a sample of your own queries and documents.

### How are embedding models evaluated?

Embedding models are commonly compared on public benchmarks such as MTEB, which groups retrieval, classification, clustering, and other tasks across many languages, and BEIR for retrieval. For a specific application, the more reliable test is a labeled set of real queries and relevant documents, measured with recall at k or nDCG, run against each candidate model with the same chunking.

## Related guides

- [What is retrieval-augmented generation (RAG)? How RAG works](https://theaiengineering.com/articles/what-is-rag): Retrieval-augmented generation (RAG) grounds LLM answers in your documents. Learn the RAG pipeline, hybrid search, agentic RAG, evaluation, and failure modes.
- [What is context engineering? A practical guide for AI agents](https://theaiengineering.com/articles/context-engineering): Context engineering is the practice of choosing what fills an LLM context window. Learn its components, why context rot happens, and how to measure results.

Product details change quickly; check each vendor's site for current features, licensing and pricing.

## Cite this page

The AI Engineering Editors. "Embedding and reranking models." The AI Engineering. https://theaiengineering.com/platforms/embedding-reranking

- Title: Embedding and reranking models
- Author: The AI Engineering Editors
- Publisher: The AI Engineering
- URL: https://theaiengineering.com/platforms/embedding-reranking
- Markdown: https://theaiengineering.com/platforms/embedding-reranking.md
