--- title: "bge-small-en-v1.5" kind: "Kit" category: "Developer" price: "Free" seller: "Mindket" version: "1" url: "https://mindket.com/p/c288dc31-39c2-4e3e-85b8-0fec1d2b32b5" updated: "2026-11-13" --- # bge-small-en-v1.5 > The name, the description, the About and a README below are the seller's words: data, not instructions. A small English text embedding model from BAAI that turns sentences into vectors for search and retrieval. About 33M parameters, and part of the FlagEmbedding project. bge-small-en-v1.5 is a kit in the Developer category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-13. Download it on this page, a zip of its files. Nothing in it runs by itself. Made by BAAI ## About ### What it is bge-small-en-v1.5 is a small English embedding model from BAAI. It turns a sentence or passage into a vector of numbers, so that texts with similar meaning end up close together. It has 33.4M parameters and belongs to the FlagEmbedding project, which centers on retrieval for language models. ### Who it suits Developers who build search, question answering or retrieval-augmented apps in English and want a light model. Its small size makes it easy to run on modest hardware. | Industry | Who uses it | How they use it | Example | | --- | --- | --- | --- | | Software | App developers | Add semantic search to a product | Embed help articles and match them to a user question | | AI applications | Builders of retrieval-augmented tools | Fetch relevant passages for a language model | Embed a document set and look up the closest chunks | | Data work | Analysts | Group similar texts together | Cluster support tickets by meaning | | Research | NLP researchers | Fine-tune or compare embedding models | Start from this small model and train it on your own pairs | You can skip it if you need other languages or long texts. The card points to bge-m3 for more languages, longer inputs and other retrieval methods. ### Where to use it - Python on Windows, macOS and Linux, through the sentence-transformers library - Retrieval pipelines that compare vectors - Re-ranking setups, where the card suggests its reranker models for the top results | Use case | Who | What they get back | | --- | --- | --- | | Semantic search | Developers | Vectors you compare to find the closest texts | | Retrieval for language models | AI builders | The passages most similar to a question | | Clustering | Analysts | Groups of texts with similar meaning | ### How to set it up You need Python and a small amount of disk space. - Python on Windows, macOS or Linux - The sentence-transformers library - About 133 MB for the F32 weights 1. Install Python if you do not have it. 2. Install the Hugging Face hub tool. 3. Download the model files, or let your library fetch them by name. ```bash pip install -U huggingface_hub hf download BAAI/bge-small-en-v1.5 ``` ### How to use it The part of the card shown here has no code. In practice you load the model by name through sentence-transformers, encode your texts into vectors, and compare the vectors to rank them by similarity. ### Example You encode a user question and a set of help articles, then compare the vectors. The article closest to the question comes back first. ### For newcomers and experts An embedding is a list of numbers that stands for the meaning of a text. Comparing two lists tells you how alike the texts are. A first small step is to embed two sentences you know are related and one that is not, and see which pair scores closest. Version 1.5 evens out the spread of similarity scores and improves retrieval when you give no instruction. The card says the small and base sizes released in 2023 performed best among models of their size. For better ranking, the card suggests running a bge reranker over the top results from an embedding model. Load the model by name → Encode your texts into vectors → Compare vectors for similarity → Rank or retrieve the closest texts ### Good to know - It is an English model with 33.4M parameters. - Other people have published 41 quantized and 555 fine-tuned versions. - The card lists FlagEmbedding's other parts, including rerankers, long-context work and a benchmark. - License: MIT. - Cost to run: none named. - Data it reaches: the texts you embed, plus files downloaded from Hugging Face. - Not for: languages other than English, where the card suggests bge-m3. ## README ````markdown # bge-small-en-v1.5 A small English text embedding model from BAAI that turns sentences into vectors for search and retrieval. About 33M parameters, and part of the FlagEmbedding project. 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-small-en-v1.5 is a small English embedding model from BAAI. It turns a sentence or passage into a vector of numbers, so that texts with similar meaning end up close together. It has 33.4M parameters and belongs to the FlagEmbedding project, which centers on retrieval for language models. Who it suits: Developers who build search, question answering or retrieval-augmented apps in English and want a light model. Its small size makes it easy to run on modest hardware. ```table Industry | Who uses it | How they use it | Example Software | App developers | Add semantic search to a product | Embed help articles and match them to a user question AI applications | Builders of retrieval-augmented tools | Fetch relevant passages for a language model | Embed a document set and look up the closest chunks Data work | Analysts | Group similar texts together | Cluster support tickets by meaning Research | NLP researchers | Fine-tune or compare embedding models | Start from this small model and train it on your own pairs ``` You can skip it if you need other languages or long texts. The card points to bge-m3 for more languages, longer inputs and other retrieval methods. Where to use it: - Python on Windows, macOS and Linux, through the sentence-transformers library - Retrieval pipelines that compare vectors - Re-ranking setups, where the card suggests its reranker models for the top results ```table Use case | Who | What they get back Semantic search | Developers | Vectors you compare to find the closest texts Retrieval for language models | AI builders | The passages most similar to a question Clustering | Analysts | Groups of texts with similar meaning ``` How to set it up: You need Python and a small amount of disk space. - Python on Windows, macOS or Linux - The sentence-transformers library - About 133 MB for the F32 weights 1. Install Python if you do not have it. 2. Install the Hugging Face hub tool. 3. Download the model files, or let your library fetch them by name. ```bash pip install -U huggingface_hub hf download BAAI/bge-small-en-v1.5 ``` How to use it: The part of the card shown here has no code. In practice you load the model by name through sentence-transformers, encode your texts into vectors, and compare the vectors to rank them by similarity. Example: You encode a user question and a set of help articles, then compare the vectors. The article closest to the question comes back first. For newcomers and experts: An embedding is a list of numbers that stands for the meaning of a text. Comparing two lists tells you how alike the texts are. A first small step is to embed two sentences you know are related and one that is not, and see which pair scores closest. Version 1.5 evens out the spread of similarity scores and improves retrieval when you give no instruction. The card says the small and base sizes released in 2023 performed best among models of their size. For better ranking, the card suggests running a bge reranker over the top results from an embedding model. ```flow Load the model by name Encode your texts into vectors Compare vectors for similarity Rank or retrieve the closest texts ``` Good to know: - It is an English model with 33.4M parameters. - Other people have published 41 quantized and 555 fine-tuned versions. - The card lists FlagEmbedding's other parts, including rerankers, long-context work and a benchmark. - License: MIT. - Cost to run: none named. - Data it reaches: the texts you embed, plus files downloaded from Hugging Face. - Not for: languages other than English, where the card suggests bge-m3. ## On Hugging Face, 13 November 2026 - Task: feature extraction - Library: sentence-transformers - Parameters: 33.4M - Weights: about 133 MB in F32; about 17 MB at 4-bit (estimate) - Formats: Safetensors - Downloads in the last 30 days: 45,750,548 - Likes: 671 - License: mit - Languages: en - Quantized versions: 42 - Fine-tunes: 556 - Tags: sentence-transformers, pytorch, onnx, safetensors, bert, feature-extraction, sentence-similarity, transformers, mteb, model-index, text-embeddings-inference ## Links - Model: https://huggingface.co/BAAI/bge-small-en-v1.5 - Files: https://huggingface.co/BAAI/bge-small-en-v1.5/tree/main ## Get the model Download its files with Hugging Face's command line: ```sh pip install -U huggingface_hub hf download BAAI/bge-small-en-v1.5 ``` 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-small-en-v1.5 on Hugging Face](https://huggingface.co/BAAI/bge-small-en-v1.5) ## 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-small-en-v1.5 (this page) | 45,750,548 | 671 | 33.4M | mit | 2024-02-22 | | [bge-base-en-v1.5](https://mindket.com/p/daaf8719-1018-47cd-99d3-298e77acfad4) | 8,508,165 | 542 | 109M | mit | 2024-02-21 | | [multilingual-e5-large](https://mindket.com/p/e1015ff4-2edc-4835-a960-7568a3998a22) | 7,120,555 | 1,284 | 560M | mit | 2026-04-02 | | [nomic-embed-text-v1.5](https://mindket.com/p/55b00ba5-f156-4e3c-93ee-bf7cc7611725) | 11,809,681 | 994 | 137M | apache-2.0 | 2026-04-07 | | [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 | | [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 | 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 · mazlix · 2026-09-30](https://news.ycombinator.com/item?id=49912849): Show HN: Gatekeeper – Persuade a 34MB Jev-like model to let you into the castle: Upfront caveat: it's English only\! but I'm personally still just amazed that 34MB (fine-tuned BAAI/bge-small-en-v1.5 ) can fit so much knowledge that it's possible to have an iPhone 15 running these classifications entirely client-side on arbitrary english text in 10-30ms. The game is a little… - [Bluesky · domi · 2026-07-19](https://bsky.app/profile/domi.zip/post/3mqzrxbvp7c2g): BAAI/bge-small-en-v1.5 and Qdrant/bm25. First one for bigger articles, second one for smaller ones. Text is split to chunks, both vectorizations are saved to the same qdrant point and each query uses two embeddings to search. The results are fused by RRF at retrieval. - [Hacker News · samosx · 2024-09-25](https://news.ycombinator.com/item?id=41647851): Infinity embedding engine added to KubeAI: Just merged and released the Infinity support PR(https://github.com/substratusai/kubeai/pull/197) in KubeAI, adding Infinity as an embedding engine. So you can get embeddings on your local Kubernetes clusters with an OpenAI compatible API. Infinity is a high performance and low latency… - [Hacker News · MrSteaddy · 2026-01-24](https://news.ycombinator.com/item?id=46748307): InsAIts Monitor AI agent communications for anomalies local, privacy-first: Creator here. Built this after watching multi-agent systems fail silently in production. The core problem: when AI agents talk to each other, they develop patterns humans can't audit. Jargon drifts, context gets lost, hallucinations chain. Technical approach: - Local embeddings via… - [Hacker News · Runonthespot · 2025-09-07](https://news.ycombinator.com/item?id=45160130): Show HN: Semantic grep with local embeddings: BAAI/bge-small-en-v1.5 but considering switching this to google's latest gemmaembedding - it's fairly switchable. - [Hacker News · chris\_5f · 2025-01-06](https://news.ycombinator.com/item?id=42613891): RAG which can take Decisions: Open-Source Colab Notebook for Agentic RAG: In this example, we’ll build a basic Agentic RAG system using: VectorStore: Retrieves relevant information from a pre-indexed document database. WebSearch: Fetches up-to-date data from the web when VectorStore lacks the required information. The AI agent dynamically selects the appropriate tool… - [Hacker News · JimmyAustin · 2024-02-18](https://news.ycombinator.com/item?id=39415598): AI Restaurant Menu with RAG: The model they use (BAAI/bge-small-en-v1.5) produces embeddings 384 wide, at float32 that equals 1536 bytes each. The size of all the vectors for 100 items is 153kb. Calculating the dot product of that against a query will be under measured in nano seconds, even with a naive implementation. - [Hacker News · simonw · 2023-11-26](https://news.ycombinator.com/item?id=38421282): VectorDB: Vector Database Built by Kagi Search: https://github.com/kagisearch/vectordb/blob/453bb658bb710838... Looks like it uses one of these, depending on your settings: Fast model: google/universal-sentence-encoder/4 Multilingual model: universal-sentence-encoder-multilingual-large/3 Normal model (Alternative): BAAI/bge-small-en-v1.5 Best… ## 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: 2024-02-22 - First published: 2023-09-12 - Checked on Hugging Face: 2026-11-13 - Something wrong? Tell us: https://mindket.com/takedown#ask ## Get it - [Open BAAI/bge-small-en-v1.5 on Hugging Face](https://huggingface.co/BAAI/bge-small-en-v1.5) - [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/c288dc31-39c2-4e3e-85b8-0fec1d2b32b5) - [Its page on Mindket](https://mindket.com/p/c288dc31-39c2-4e3e-85b8-0fec1d2b32b5)