• Home
  • Categories
  • Pricing
  • Submit
    Built with
    Ever Works
    Ever Works

    Connect with us

    Stay Updated

    Get the latest updates and exclusive content delivered to your inbox.

    Product

    • Categories
    • Pricing
    • Help

    Clients

    • Sign In
    • Register
    • Forgot password?

    Company

    • About Us
    • Admin
    • Sitemap

    Resources

    • Blog
    • Submit
    • API Documentation
    All product names, logos, and brands are the property of their respective owners. All company, product, and service names used in this repository, related repositories, and associated websites are for identification purposes only. The use of these names, logos, and brands does not imply endorsement, affiliation, or sponsorship. This directory may include content generated by artificial intelligence.
    Copyright © 2025 Awesome Vector Databases. All rights reserved.·Terms of Service·Privacy Policy·Cookies
    Decorative pattern
    Decorative pattern
    1. Home
    2. Machine Learning Models
    3. mxbai-rerank-base-v2

    mxbai-rerank-base-v2

    A 0.5B parameter reranking model by Mixedbread AI that provides an excellent balance of speed and accuracy, supporting 100+ languages and processing up to 8K tokens with reinforcement learning training for enhanced search relevance.

    Overview

    mxbai-rerank-base-v2 is a second-generation reranking model from Mixedbread AI that combines compact size with state-of-the-art performance. It is fully open-source under the Apache 2.0 license.

    Key Features

    • Reinforcement Learning: Trained using GRPO (Guided Reinforcement Prompt Optimization) for enhanced performance
    • Multilingual: Supports 100+ languages for global applications
    • Extended Context: Handles up to 8K tokens (32K-compatible) for comprehensive document analysis
    • Fast Processing: Up to 8x faster than comparable solutions
    • Compact Size: 0.5B parameters provide optimal balance of size and performance

    Performance

    Benchmark results on standard datasets:

    • BEIR Average: 55.57 NDCG@10
    • Mr.TyDi (Multilingual): 28.56 NDCG@10
    • Chinese Datasets: 83.70 NDCG@10
    • Code Search: 31.73 NDCG@10

    Integration

    The rerank model can be added to the end of an existing keyword-based search workflow (Elasticsearch, OpenSearch, Solr), allowing semantic relevance to be incorporated without changing existing infrastructure.

    Use Cases

    • Improving search relevance in RAG applications
    • Re-ranking results from BM25 or vector search
    • Multi-stage retrieval pipelines
    • Code search and technical documentation
    • Multilingual search applications

    Pricing

    Free and open-source under Apache 2.0 license. Also available via Mixedbread API with usage-based pricing.

    Surveys

    Loading more......

    Information

    Websitewww.mixedbread.com
    PublishedMar 20, 2026

    Categories

    1 Item
    Machine Learning Models

    Tags

    3 Items
    #reranker#multilingual#open-source

    Similar Products

    6 result(s)

    Cohere Rerank v3.5

    State-of-the-art foundational model for ranking with 4096 context length and multilingual support for 100+ languages. Offers exceptional performance on BEIR benchmarks and specialized domains including finance, e-commerce, and enterprise search.

    Featured

    Qwen3 Embedding

    Multilingual embedding model supporting over 100 languages and ranking #1 on MTEB multilingual leaderboard. Offers flexible model sizes from 0.6B to 8B parameters with user-defined instructions.

    Featured

    Nomic Embed Text

    First fully reproducible open-source text embedding model with 8,192 context length. v2 introduces Mixture-of-Experts architecture for multilingual embeddings. Outperforms OpenAI models on benchmarks. This is an OSS model under Apache 2.0 license.

    Featured

    Jina Reranker v2

    Transformer-based cross-encoder model fine-tuned for text reranking with Flash Attention 2 architecture. Features multilingual support for 100+ languages, function-calling capabilities, code search, and 6x speedup over v1 with only 278M parameters.

    BGE-M3

    A versatile multilingual text embedding model from BAAI that supports 100+ languages and can handle inputs up to 8192 tokens. BGE-M3 is unique in supporting three retrieval methods simultaneously: dense retrieval, multi-vector retrieval, and sparse retrieval.

    gte-Qwen2-1.5B-instruct

    A state-of-the-art multilingual text embedding model from Alibaba's GTE (General Text Embedding) series, built on the Qwen2-1.5B LLM. The model supports up to 8192 tokens and incorporates bidirectional attention mechanisms for enhanced contextual understanding across diverse domains.