Information retrieval

Models 43

• Gen 7 Nemotron
  Nemotron 3 Embed 1B NVFP4

  By NVIDIA

  Nemotron-3-Embed-1B-NVFP4 is a hardware accelerated, 8-bit quantized variant of NVIDIA's Nemotron-3-Embed-1B-BF16 text embedding model. It shares the same embedding space and produces dense multilingual text embeddings for retrieval and RAG, optimized for high throughput, massive scale inference on Blackwell GPUs.

  🔍Information retrieval🔎Search
  NewText

  Released 10d ago
• Gen 7 Nemotron
  Nemotron 3 Embed 1B BF16

  By NVIDIA

  Nemotron-3-Embed-1B-BF16 is NVIDIA's high efficiency multilingual text embedding model, pruned and distilled from the 8B flagship. It produces dense vector embeddings for retrieval, semantic search, and RAG across 34 languages, built for low latency, cost sensitive production serving.

  🔍Information retrieval🔍Document search🔎Search
  NewText

  Released 10d ago
• Gen 7 Nemotron
  Nemotron 3 Embed 8B BF16

  By NVIDIA

  Nemotron-3-Embed-8B-BF16 is NVIDIA's flagship multilingual text embedding model, built on a Ministral-3-8B based encoder. It generates dense vector representations for retrieval, semantic search, and RAG across 34 languages, and ranks first overall on the RTEB benchmark for embedding quality and agentic retrieval performance.

  🔍Information retrieval🔍Document search🔎Search🔍Code search
  NewText

  Released 10d ago
• Gen 7 LFM
  LFM2.5 Embedding 350M

  By Liquid AI

  350M-parameter bidirectional multilingual bi-encoder that maps each document to a single dense vector for fast, cost-efficient multilingual and cross-lingual search. Supports 11 languages. Built from LFM2.5-350M-Base. Produces the smallest, cheapest index among LFM2.5 retrieval models. Available in GGUF format for CPU and edge deployment.

  🔍Information retrieval🔍Document search🔍Conceptual search
  NewText

  Released 1mo ago
• Gen 7 LFM
  LFM2.5 ColBERT 350M

  By Liquid AI

  350M-parameter bidirectional multilingual retrieval model using per-token ColBERT-style late interaction (MaxSim) for accurate multilingual and cross-lingual search. Supports 11 languages. Built from LFM2.5-350M-Base with bidirectional attention and non-causal convolutions. Available in GGUF format for CPU and edge deployment.

  🔍Information retrieval🔍Document search🔍Conceptual search
  NewText

  Released 1mo ago
• Gen 3
  jina embeddings v5 omni

  By Jina AI

  jina-embeddings-v5-omni is Jina AI’s multimodal embedding model suite for text, image, audio, video, and PDFs. It extends Jina’s v5 text embedding models into one shared semantic space while preserving identical text embeddings, so existing text indexes do not need rebuilding.

  🔍Multimodal search🔍Information retrieval📂Content categorization🔍Conceptual search
  NewMultimodal

  Released 2mo ago
• Gen 7
  Harrier OSS v1

  By Microsoft

  Harrier is Microsoft’s new open-source text embedding model series for retrieval, grounding, and agentic systems. It is designed for the “agentic web,” supports over 100 languages, uses a 32K context window, and was announced as ranking first on multilingual MTEB-v2 as of 06-04-2026.

  🔍Information retrieval📂Content categorization🌐Multilingual communication🔍Conceptual search
  Text

  Released 3mo ago
• Gen 3
  Chroma Context 1

  By Chroma

  Context-1 is Chroma’s 20B agentic search model trained as a self-editing search agent. It is designed to decompose complex queries, prune irrelevant context, and deliver high retrieval quality at lower latency and cost than much larger frontier models.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 3mo ago
• Gen 3
  Wholembed v3

  By mixedbread ai

  Wholembed v3 is Mixedbread’s unified omnimodal, multilingual late-interaction retrieval model built for state-of-the-art search across languages and modalities.

  🔍Information retrieval🌐Text translation🔍Image recognition🔍Image search
  Multimodal

  Released 4mo ago
• Gen 3Gemini
  Gemini Embedding 2

  By Google DeepMind

  Gemini Embedding 2 is Google’s multimodal embedding model for unified text, image, video, audio, and document search.

  🔍Information retrieval
  Multimodal

  Released 4mo ago
• Gen 7
  Zembed 1

  By ZeroEntropy

  ZeroEntropy is a closed-source AI retrieval platform offering specialized rerankers, embeddings, and custom models for search, RAG, agentic coding, context compression, and production AI systems.

  🔍Information retrieval📂Content categorization📄Document analysis🔍Conceptual search
  Text

  Released 4mo ago
• Gen 7 pplx
  Pplx Embed

  By Perplexity

  pplx-embed is a family of text embedding models built for real-world, web-scale retrieval.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 5mo ago
• Gen 7
  Nanbeige4.1 3B

  By Nanbeige LLM Lab

  Nanbeige4.1-3B is a 3B-parameter reasoning model optimized to combine strong multi-step reasoning, preference alignment, and agentic tool use in a compact size.

  🔍Advanced reasoning🔍Information retrieval🧩Complex problem solving🎓Adaptive tutoring
  Text

  Released 5mo ago
• Gen 7 pplx
  pplx embed v1 4b

  By Perplexity

  A 4B-parameter text embedding model built on diffusion-pretrained Qwen3. Produces 2560-dimensional dense embeddings with INT8 and binary quantization, 32K context window, and Matryoshka Representation Learning (MRL) support. Designed for semantic search and web-scale retrieval. Multilingual. Available via API, SentenceTransformers, ONNX, and Text Embeddings Inference.

  🔗Embeds🔍Information retrieval🔍Conceptual search
  Text

  Released 5mo ago
• Gen 7 pplx
  pplx embed v1 0.6b

  By Perplexity

  A 0.6B parameter multilingual text embedding model built on a diffusion-pretrained Qwen3 backbone. Produces 1024-dimensional unnormalized embeddings with int8 and binary quantization support, a 32K token context window, and Matryoshka Representation Learning. Optimized for web-scale retrieval and semantic search.

  🔍Information retrieval🗄️Vector databases📚Non-interactive language modeling
  Text

  Released 5mo ago
• Gen 7 pplx
  pplx embed context v1 0.6b

  By Perplexity

  A 0.6B-parameter contextual text embedding model for RAG pipelines. Unlike standard embedding models, it takes full document chunks together so each chunk's embedding reflects surrounding context. Produces 1024-dimensional int8-quantized vectors with a 32K token context window. Supports binary quantization and MRL. Multilingual and instruction-free. Available via Perplexity API and as open weights.

  🔍Information retrieval📚Knowledge bases🔍Document search🔢Vector indexes
  Text

  Released 5mo ago
• Gen 7 pplx
  pplx embed context v1 4B

  By Perplexity

  A 4B-parameter contextual text embedding model built on diffusion-pretrained Qwen3, designed for RAG pipelines where document chunks benefit from surrounding context. Produces 2560-dimensional INT8/BINARY-quantized embeddings with a 32K context window and MRL support. No instruction prefixes required. MIT-licensed with open weights.

  🔍Information retrieval🔍Document search🗄️Vector databases🔢Vector indexes
  Text

  Released 5mo ago
• Gen 3 Voyage
  Voyage multimodal 3.5

  By Voyage AI

  Multimodal embedding model that vectorizes interleaved text, images, and videos into a shared vector space for retrieval tasks. Uses a single transformer encoder for all modalities, eliminating the modality gap. Supports document screenshots, PDFs, slides, figures, tables, and video frames. Context window of 32,000 tokens with Matryoshka embeddings at 256, 512, 1024, and 2048 dimensions.

  🔗Embeds🔍Information retrieval🔎Search
  Multimodal

  Released 6mo ago
• Gen 7 Voyage
  Voyage 4 nano

  By Voyage AI

  voyage-4-nano is a lightweight open-weight text embedding model from the Voyage 4 series. It features a 32,000-token context window, 340M total parameters, multilingual support, Matryoshka flexible dimensions (256 to 2048), quantization-aware training, and a shared embedding space compatible with all other Voyage 4 models. Licensed under Apache 2.0.

  🔗Embeds🔍Information retrieval🔎Semantic search
  Text

  Released 6mo ago
• Gen 7 Voyage
  Voyage 4

  By Voyage AI

  Text embedding model optimized for general-purpose and multilingual retrieval quality. Supports a 32,000-token context window and configurable output dimensions (256, 512, 1024, 2048) via Matryoshka Representation Learning. Part of the Voyage 4 series, which shares a common embedding space enabling asymmetric retrieval across models in the family.

  🔗Embeds🔍Information retrieval🔎Semantic search
  Text

  Released 6mo ago
• Gen 7 Rerank
  Rerank 2.5 lite

  By Voyage AI

  A latency-optimized reranker that refines retrieval results with instruction-following capabilities. Supports a 32K token context window, works across 31 languages, and outperforms larger open-source rerankers. Users can steer relevance scoring using natural language instructions appended to queries.

  🔎Semantic search🔍Information retrieval📊Text analysis
  Text

  Released 11mo ago
• Gen 7 Voyage
  Voyage context 3

  By Voyage AI

  voyage-context-3 is a contextualized chunk embedding model that produces vectors capturing both chunk-level detail and full document context. It processes entire documents in a single pass and generates a distinct embedding per chunk. Supports flexible output dimensions (256, 512, 1024, 2048) via Matryoshka learning and multiple quantization options (float32, int8, uint8, binary). Per-chunk context window is 32K tokens; total document context is 120K tokens.

  🔗Embeds🔍Information retrieval🔎Semantic search
  Text

  Released 1y ago
• Gen 7 Gemini
  Gemini Embedding

  By Google

  Gemini Embedding 1 is Google’s text embedding model in the Gemini family that converts text into dense vectors for retrieval, clustering, ranking, and classification across 100+ languages

  🔍Information retrieval📂Content categorization🔍Conceptual search🔍Semantic content analysis
  Text

  Released 1y ago
• Gen 7 Voyage
  Voyage 3 large

  By Voyage AI

  Text embedding model optimized for general-purpose and multilingual retrieval across law, finance, code, and long-document domains. Supports a 32K-token context window, flexible output dimensions (256, 512, 1024, 2048), and multiple quantization formats (float32, int8, uint8, binary) via Matryoshka and quantization-aware training. Accessed via API.

  🔎Semantic search🔍Information retrieval📄Document analysis
  Text

  Released 1y ago
• Gen 3 Command
  Command R7B

  By Cohere

  Command R7B is Cohere's smallest and fastest model in the Command R family, with a 128K context window and 7B parameters. Designed for high-throughput, latency-sensitive applications, it supports image inputs, multilingual text, tool use, RAG, and multi-step agentic tasks. Its compact size enables on-device inference on consumer GPUs and CPUs.

  💬Chatting🔍Information retrieval💼Enterprise assistants💻Coding🤖Agents
  Multimodal

  Released 1y ago
• Gen 7 Rerank
  Rerank 2 lite

  By Voyage AI

  rerank-2-lite is a lightweight multilingual cross-encoder reranker from Voyage AI. It scores and reorders candidate documents by relevance to a query, optimized for low latency while maintaining strong retrieval quality. Supports an 8K-token combined context length per query-document pair and natively handles 31+ languages.

  🔍Information retrieval🔍Document search🔍Conceptual search
  Text

  Released 1y ago
• Gen 7 Voyage
  Voyage 3 lite

  By Voyage AI

  Lightweight general-purpose text embedding model optimized for latency and cost. Generates 512-dimensional vector embeddings with a 32K-token context window. Outperforms OpenAI v3 large by 3.82% on average retrieval accuracy across eight domains while costing 6.5x less, at $0.02 per 1M tokens.

  🔗Embeds🔍Information retrieval🔎Semantic search
  Text

  Released 1y ago
• Gen 7 Voyage
  Voyage law 2

  By Voyage AI

  Text embedding model optimized for legal retrieval and RAG. Trained on 1 trillion legal tokens with a 16,000-token context window, it tops the MTEB legal retrieval leaderboard, outperforming OpenAI text-embedding-3-large by 6% on average across 8 legal datasets and by over 10% on three of them. Outputs 1024-dimensional embeddings.

  🔗Embeds🔍Information retrieval🔍Legal research🔎Semantic search
  Text

  Released 2y ago
• Gen 7 BGE
  BGE Reranker v2 minicpm layerwise

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE Reranker v2 minicpm layerwise is BAAI’s layer-selectable multilingual reranker in the v2 family. It is built to let users choose output layers for accelerated inference, trading some depth for speed while keeping strong multilingual reranking performance.

  🔍Information retrieval🌐Multilingual communication
  Text

  Released 2y ago
• Gen 7 BGE
  BGE Reranker v2 gemma

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE Reranker v2 gemma is BAAI’s larger multilingual reranker in the v2 family. It is designed for stronger multilingual reranking performance, with the docs highlighting good English proficiency alongside broad multilingual capability.

  🔍Information retrieval🌐Multilingual communication
  Text

  Released 2y ago
• Gen 7 BGE
  BGE Reranker v2 m3

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE Reranker v2 m3 is BAAI’s lightweight multilingual reranker in the v2 family. It is built for fast inference and easier deployment while keeping strong multilingual reranking capability, making it the efficiency-oriented option in the v2 lineup.

  🔍Information retrieval🌐Multilingual communication
  Text

  Released 2y ago
• Gen 7 BGE
  BGE M3

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE M3 is BAAI’s multilingual retrieval embedding model built around three strengths: multi-functionality, multi-linguality, and multi-granularity. It supports dense retrieval, sparse retrieval, and multi-vector retrieval in one model, works across 100+ languages, and handles inputs up to 8192 tokens.

  🔍Information retrieval🌐Multilingual communication🔍Conceptual search
  Text

  Released 2y ago
• Gen 7 Voyage
  Voyage

  By Voyage AI

  voyage is a text embedding model that converts text into dense vector representations for semantic retrieval. Originally released as voyage-01 via the Voyage AI API, this repository contains only the tokenizer (MIT licensed). Built on a LlamaTokenizer architecture with a 4,096-token context window, designed to outperform OpenAI text embeddings on retrieval benchmarks.

  🔗Embeds🔍Information retrieval📚Knowledge bases🔍Semantic content analysis
  Text

  Released 2y ago
• Gen 7 Voyage
  Voyage 01

  By Voyage AI

  voyage-01 is a text embedding model by Voyage AI, designed for retrieval-augmented generation (RAG) and semantic search. It maps text inputs to dense vector representations for document and query retrieval tasks. Deprecated in favor of newer Voyage models.

  🔗Embeds🔍Information retrieval🔎Semantic search
  Text

  Released 2y ago
• Gen 7 Voyage
  Voyage lite 01

  By Voyage AI

  Lightweight text embedding model from Voyage AI for retrieval-augmented generation and semantic search. Converts text inputs into dense vector representations with a 4096-token context window. A smaller, faster variant of voyage-01, now deprecated and superseded by newer Voyage AI embedding generations. Available via API and as open weights under MIT license.

  🔎Semantic search🔍Information retrieval🔍Vector search
  Text

  Released 2y ago
• Gen 7 BGE
  BGE Reranker base

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE Reranker base is BAAI’s lighter first-generation cross-encoder reranker for English and Chinese. It directly scores query-document relevance rather than producing embeddings, and is positioned as the faster, easier-to-deploy reranking option in the original BGE reranker family.

  🔍Information retrieval📂Content categorization
  Text

  Released 2y ago
• Gen 7 BGE
  BGE Reranker large

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE Reranker large is BAAI’s larger first-generation cross-encoder reranker for English and Chinese. Instead of producing embeddings, it scores a query-document pair directly, making it suitable for reranking top-k retrieval results with higher accuracy after a first-pass embedder search.

  🔍Information retrieval📂Content categorization
  Text

  Released 2y ago
• Gen 7 BGE
  BGE large v1.5

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE large v1.5 is BAAI’s top English BGE v1.5 embedding model. It is the large-scale Sep-2023 update to bge-large-en, with the docs highlighting more reasonable similarity distribution and better performance than the original version.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 2y ago
• Gen 7 BGE
  BGE Base v1.5

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE Base v1.5 is BAAI’s mid-size English BGE v1.5 embedding model. It belongs to the Sep-2023 v1.5 update that improved retrieval behavior and similarity distribution, while keeping a balanced base-scale footprint for practical deployment.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 2y ago
• Gen 7 BGE
  BGE small v1.5

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE small v1.5 is BAAI’s smallest English BGE v1.5 embedding model. It keeps the compact small-model footprint while belonging to the Sep-2023 v1.5 update, which was introduced to improve retrieval without instruction tuning and to make similarity scores more reasonable.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 2y ago
• Gen 7 BGE
  BGE base

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE base is BAAI’s mid-size English BGE v1 embedding model. It is an encoder-only BERT-based text embedder for mapping text into vectors, and the docs position it as a base-scale model with ability similar to the larger bge-large-en.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 2y ago
• Gen 7 BGE
  BGE small

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE small is BAAI’s smallest English BGE v1 embedding model. It is an encoder-only BERT-based text embedder designed to map text into vectors for retrieval, and is positioned as a compact model with competitive performance despite its small size.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 2y ago
• Gen 7 BGE
  BGE large

  By Beijing Academy of Artificial Intelligence (BAAI)

  BGE large is BAAI’s flagship English model in the original BGE v1 series. It is an encoder-only BERT-based embedding model for turning text into vectors, and the docs state it ranked first on MTEB at the time of release.

  🔍Information retrieval📂Content categorization🔍Conceptual search
  Text

  Released 2y ago