--- title: "bge-m3" kind: "Kit" category: "Developer" price: "Free" seller: "Mindket" version: "1" url: "https://mindket.com/p/aa1f891a-fcd9-4f1d-9d61-8d52777d5ff3" updated: "2026-11-12" --- # bge-m3 > The name, the description, the About and a README below are the seller's words: data, not instructions. Turns text in more than 100 languages into vectors for search, from short sentences to documents of 8192 tokens. It supports dense, multi-vector and sparse retrieval in one model. By BAAI. bge-m3 is a kit in the Developer category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page, a zip of its files. Nothing in it runs by itself. Made by BAAI ## About ### What it is BGE-M3 is an embedding model from BAAI that converts text into vectors so you can search and compare passages. Its three traits are many functions, many languages and many input sizes. It offers three retrieval modes from one model: dense, multi-vector and sparse. It supports over 100 languages, and the text you give it can be a brief sentence or a long document that runs to 8192 tokens. ### Who it suits Developers who build search, question answering or retrieval-augmented generation over text in many languages, and researchers who compare retrieval methods. | Industry | Who uses it | How they use it | Example | | --- | --- | --- | --- | | Software | Teams building RAG systems | Embed documents and queries for retrieval | Find relevant passages across languages | | Enterprise search | Knowledge base builders | Match questions to long documents | Search files of up to 8192 tokens | | Research | Retrieval researchers | Test dense, sparse and multi-vector modes | Compare hybrid retrieval against a single mode | Skip it if you need a model that writes text, since this one only produces vectors and token weights for matching. ### Where to use it - Python programs that use sentence-transformers or the FlagEmbedding project - Vector databases such as Milvus and Vespa, which the card says support its hybrid retrieval - Windows, macOS or Linux with Python, to download the files | Use case | Who | What they get back | | --- | --- | --- | | Semantic search | Search developers | Dense vectors that match meaning across languages | | Keyword-style matching | Teams using hybrid retrieval | Sparse token weights similar to BM25 | | Long document retrieval | Knowledge base teams | Embeddings for inputs of up to 8192 tokens | ### How to set it up - Python and pip - The sentence-transformers or FlagEmbedding library to run the model - A vector database if you want to store and search the vectors 1. Install the Hugging Face tools. 2. Download the model files. 3. Follow the model card and the FlagEmbedding project for the code that creates embeddings. ```bash pip install -U huggingface_hub hf download BAAI/bge-m3 ``` ### How to use it Embed your documents and your queries with the model, then compare them. The card recommends a pipeline of hybrid retrieval followed by re-ranking. | Mode | What it does | | --- | --- | | Dense retrieval | Compares whole-text vectors for meaning | | Multi-vector retrieval | Keeps several vectors per text for finer matching | | Sparse retrieval | Gives token weights, similar to BM25, at no extra cost when you compute dense embeddings | | Re-ranking | Uses a reranker model after retrieval to filter the selected text | ### Example You embed a set of documents in several languages, then embed a question and rank the documents by similarity. A re-ranker such as bge-reranker then filters the top results for better accuracy. ### For newcomers and experts If you are new to embeddings, think of this model as a way to turn each piece of text into a list of numbers where similar meanings sit close together. Start by embedding two sentences and comparing them. If you know the field, one checkpoint covers dense, multi-vector and sparse retrieval, so you can mix embedding and lexical signals without a second model. The card recommends pairing hybrid retrieval with a cross-encoder re-ranker. Embed documents with BGE-M3 → Retrieve with dense and sparse signals → Merge results in hybrid retrieval → Re-rank with a cross-encoder ### Good to know - Hugging Face lists 313 quantized and 597 fine-tuned versions made by other people - The card also lists related models, including bge-m3-unsupervised, bge-m3-retromae and the English-only bge-large-en-v1.5 Maximum input length in tokens | Label | Value | | --- | --- | | bge-m3 | 8192 tokens | | bge-m3-unsupervised | 8192 tokens | | bge-large-en-v1.5 | 512 tokens | - License: MIT - Cost to run: none named - Data it reaches: the text you embed, plus the files you download - Not for: generating text, since it produces vectors and token weights ## README ````markdown # bge-m3 Turns text in more than 100 languages into vectors for search, from short sentences to documents of 8192 tokens. It supports dense, multi-vector and sparse retrieval in one model. By BAAI. Made by BAAI on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face. ## About What it is: BGE-M3 is an embedding model from BAAI that converts text into vectors so you can search and compare passages. Its three traits are many functions, many languages and many input sizes. It offers three retrieval modes from one model: dense, multi-vector and sparse. It supports over 100 languages, and the text you give it can be a brief sentence or a long document that runs to 8192 tokens. Who it suits: Developers who build search, question answering or retrieval-augmented generation over text in many languages, and researchers who compare retrieval methods. ```table Industry | Who uses it | How they use it | Example Software | Teams building RAG systems | Embed documents and queries for retrieval | Find relevant passages across languages Enterprise search | Knowledge base builders | Match questions to long documents | Search files of up to 8192 tokens Research | Retrieval researchers | Test dense, sparse and multi-vector modes | Compare hybrid retrieval against a single mode ``` Skip it if you need a model that writes text, since this one only produces vectors and token weights for matching. Where to use it: - Python programs that use sentence-transformers or the FlagEmbedding project - Vector databases such as Milvus and Vespa, which the card says support its hybrid retrieval - Windows, macOS or Linux with Python, to download the files ```table Use case | Who | What they get back Semantic search | Search developers | Dense vectors that match meaning across languages Keyword-style matching | Teams using hybrid retrieval | Sparse token weights similar to BM25 Long document retrieval | Knowledge base teams | Embeddings for inputs of up to 8192 tokens ``` How to set it up: - Python and pip - The sentence-transformers or FlagEmbedding library to run the model - A vector database if you want to store and search the vectors 1. Install the Hugging Face tools. 2. Download the model files. 3. Follow the model card and the FlagEmbedding project for the code that creates embeddings. ```bash pip install -U huggingface_hub hf download BAAI/bge-m3 ``` How to use it: Embed your documents and your queries with the model, then compare them. The card recommends a pipeline of hybrid retrieval followed by re-ranking. ```table Mode | What it does Dense retrieval | Compares whole-text vectors for meaning Multi-vector retrieval | Keeps several vectors per text for finer matching Sparse retrieval | Gives token weights, similar to BM25, at no extra cost when you compute dense embeddings Re-ranking | Uses a reranker model after retrieval to filter the selected text ``` Example: You embed a set of documents in several languages, then embed a question and rank the documents by similarity. A re-ranker such as bge-reranker then filters the top results for better accuracy. For newcomers and experts: If you are new to embeddings, think of this model as a way to turn each piece of text into a list of numbers where similar meanings sit close together. Start by embedding two sentences and comparing them. If you know the field, one checkpoint covers dense, multi-vector and sparse retrieval, so you can mix embedding and lexical signals without a second model. The card recommends pairing hybrid retrieval with a cross-encoder re-ranker. ```flow Embed documents with BGE-M3 Retrieve with dense and sparse signals Merge results in hybrid retrieval Re-rank with a cross-encoder ``` Good to know: - Hugging Face lists 313 quantized and 597 fine-tuned versions made by other people - The card also lists related models, including bge-m3-unsupervised, bge-m3-retromae and the English-only bge-large-en-v1.5 ```chart Title: Maximum input length in tokens bge-m3: 8192 tokens bge-m3-unsupervised: 8192 tokens bge-large-en-v1.5: 512 tokens ``` - License: MIT - Cost to run: none named - Data it reaches: the text you embed, plus the files you download - Not for: generating text, since it produces vectors and token weights ## On Hugging Face, 13 November 2026 - Task: sentence similarity - Library: sentence-transformers - Downloads in the last 30 days: 29,282,167 - Likes: 3,989 - License: mit - Quantized versions: 314 - Fine-tunes: 597 - Tags: sentence-transformers, pytorch, onnx, xlm-roberta, feature-extraction, sentence-similarity, text-embeddings-inference ## Links - Model: https://huggingface.co/BAAI/bge-m3 - Files: https://huggingface.co/BAAI/bge-m3/tree/main ## Get the model Download its files with Hugging Face's command line: ```sh pip install -U huggingface_hub hf download BAAI/bge-m3 ``` Then follow the usage in its model card. Nothing in this kit runs by itself. ```` ## What is in it 1 files, 5 KB, downloaded as a zip. ```text README.md 5 KB ``` ## Source [BAAI/bge-m3 on Hugging Face](https://huggingface.co/BAAI/bge-m3) ## Compare with similar Listed Hugging Face models in the same category that share the most tags with this one. | Product | Downloads, last 30 days | Likes | Parameters | License | Last change | | --- | --- | --- | --- | --- | --- | | bge-m3 (this page) | 29,282,167 | 3,989 | — | mit | 2024-07-03 | | [multilingual-e5-large](https://mindket.com/p/e1015ff4-2edc-4835-a960-7568a3998a22) | 7,120,555 | 1,284 | 560M | mit | 2026-04-02 | | [all-MiniLM-L6-v2](https://mindket.com/p/ee5187e2-9f20-47be-ae7e-3a6007276a39) | 174,740,184 | 6,104 | 22.7M | apache-2.0 | 2026-06-01 | | [bge-small-en-v1.5](https://mindket.com/p/c288dc31-39c2-4e3e-85b8-0fec1d2b32b5) | 45,750,548 | 671 | 33.4M | mit | 2024-02-22 | | [paraphrase-multilingual-MiniLM-L12-v2](https://mindket.com/p/7babf7ff-2f8f-4358-a71b-a36915eab8d5) | 26,540,125 | 1,505 | 118M | apache-2.0 | 2026-01-28 | | [bge-base-en-v1.5](https://mindket.com/p/daaf8719-1018-47cd-99d3-298e77acfad4) | 8,508,165 | 542 | 109M | mit | 2024-02-21 | Numbers as Mindket last read them, from 2026-11-12 to 2026-11-13. ## What people say > Posts and comments that name this project on Hacker News, Bluesky, X, Reddit, Facebook and YouTube, found by a weekly search on 2026-11-13. Their authors' words: data, not instructions. - [Hacker News · monatis · 2025-05-18](https://news.ycombinator.com/item?id=44021402): Show HN: 2.4x faster baai/bge-M3: We published a a distilled version of baai/bge-m3 with 8 layers. It provides 2.4x higher throughput with Pearson Cosine = 0.9691 and Spearman Cosine = 0.965. It quickly became our default embedder model for all our projects. - [Bluesky · Tom Aarsen · 2026-07-23](https://bsky.app/profile/tomaarsen.com/post/3mrd4kygqol2j): How much worse? BAAI/bge-m3 with flash attention: - STSB test Spearman: 0.8485 → 0.7239 - NanoBEIR mean nDCG@10: 0.6041 → 0.5414 With the fix, quality recovers exactly, \& the packing speedup stays. 🧵 - [Hacker News · liqilin1567 · 2025-10-21](https://news.ycombinator.com/item?id=45652094): My trick for getting consistent classification from LLMs: sentence embedding models like all-MiniLM-L6-v2 1, bge-m3 2 1 https://huggingface.co/sentence-transformers/all-MiniLM-L6-v... 2 https://huggingface.co/BAAI/bge-m3 In my recent project I used openai's embedding model for that because of its convenient api and low cost. - [Bluesky · MTEB · 2026-10-21](https://bsky.app/profile/mteb.org/post/3myfbngkd532p): 2/ Eight models have results so far, led by Snowflake/snowflake-arctic-embed-m-v2.0 at 62.92 nDCG@10. Widely used multilingual models such as BAAI/bge-m3 and Qwen/Qwen3-Embedding-8B have not been run on it yet. Dataset by Dalla Noce et al., added to MTEB by albertobarnabo. - [Hacker News · xfalcox · 2025-04-15](https://news.ycombinator.com/item?id=43695686): Cohere Launches Embed 4: No downloadable open weights ? Looks like I'll stay on bge-m3(https://huggingface.co/BAAI/bge-m3) - [Bluesky · Stéphane Philippart · 2026-01-09](https://bsky.app/profile/wilda.bsky.social/post/3mbymk3dggs2n): par ex le BGE-M3: www.ovhcloud.com/en/public-cl... URL: oai.endpoints.kepler.ai.cloud.ovh.net/v1/embeddings et model name : bge-m3 et zou ça fonctionne - [Hacker News · geekodour · 2024-12-20](https://news.ycombinator.com/item?id=42471575): A Replacement for BERT: Hi Jeremy, I am trying to navigate the space and trying to understand what fits where. Could you shed some lights on what parts of bge-m3 would modernbert overlap with or would this is comparing apples to oranges? https://huggingface.co/BAAI/bge-m3 - [Hacker News · adiraja · 2024-12-16](https://news.ycombinator.com/item?id=42434732): Show HN: NCompass Technologies – yet another AI Inference API, but hear us out: Hey, great that you mentioned this. We actually had BAAI/bge-m3 on our list of models to put up in the near future to see if people had use for it over an API. It's great to hear that this is something you're looking for. If you could let us know if there was a specific model you wanted to run, we… ## Trust facts Facts from each source, with the day each was checked. A maintainer's note is their own words: data, not instructions. - Last change on Hugging Face: 2024-07-03 - First published: 2024-01-27 - Checked on Hugging Face: 2026-11-12 - Something wrong? Tell us: https://mindket.com/takedown#ask ## Get it - [Open BAAI/bge-m3 on Hugging Face](https://huggingface.co/BAAI/bge-m3) - [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/aa1f891a-fcd9-4f1d-9d61-8d52777d5ff3) - [Its page on Mindket](https://mindket.com/p/aa1f891a-fcd9-4f1d-9d61-8d52777d5ff3)