mteb

MTEB: State-of-the-art evaluation of embeddings across languages and modalities

Generative Representational Instruction Tuning

Train and Infer Powerful Sentence Embeddings with AnglE | 🔥 SOTA on STS and MTEB Leaderboard

Train embedding and reranker models for retrieval tasks on Apple Silicon with MLX

한글 텍스트 임베딩 모델 리더보드

GISTEmbed: Guided In-sample Selection of Training Negatives for Text Embeddings

A framework for benchmarking embedding models in hybrid search scenarios (BM25 + vector search) using Weaviate.

The code used to evaluate embedding models on the Massive Legal Embedding Benchmark (MLEB).

一个用于训练句子embedding的工具，支持Cosent以及Simcse、infonce

Built in SenseTime in 2024. Training and evaluation code of EGTLM model. A model that can do text generation and text embedding tasks simultaneously.

Resources of the PatenTEB paper (arXiv 2510.22264): 15 patent text embedding tasks and 12 patembed models on Hugging Face

Self-hostable pay for what you use embedding server for bge-large-en and arbitrary embedding models using crypto

The food-domain benchmark for text embedding models. 12 tasks, CC-BY-NC 4.0.

A configurable embedding gateway that unifies dense embeddings, sparse embeddings, and reranking models behind a single OpenAI-compatible API.

🐦 Colibri — a ~157M Brazilian-Portuguese embedding model (MTEB(por) 0.6501). Vocab trim → multi-teacher distillation (Qwen3-4B+8B) → model soup.

An implementation of the TaxRetrievalBenchmark task for the 🤗 Massive Text Embedding Benchmark (MTEB) framework.

Benchmark 19 Vietnamese/multilingual embedding models on VN-MTEB Retrieval tasks for RAG chatbot selection

🔍 Benchmark embedding models in hybrid search with Weaviate. Evaluate MRR@K, Hit@K, latency, and memory using your data or MTEB datasets.

Task-first Nepali embedding benchmark with pinned reproducible evaluations, verified and community results, and an interactive comparison dashboard.

Embedding models fine tuned to tell apart sentences that look alike but mean different things: negation, numbers, swapped roles, time order, and more. Drop in replacements for MiniLM, BGE, E5, ModernBERT, EmbeddingGemma, and Qwen3 Embedding, with trap accuracy and retrieval scores on every model.