# Qwen3-Embedding-4B by Qwen: Open-Weight Model Source: https://savrn.com/models/qwen3-embedding-4b Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Runs On What it takes to serve Qwen3-Embedding-4B (4B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them. | Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits | | --- | --- | --- | --- | --- | --- | | 16-bit | 8.0 GB | 9.7 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 8-bit | 4.0 GB | 4.8 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | | 4-bit | 2.0 GB | 2.4 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 | Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the [SAVRN Index](https://savrn.com/ai-index/pricing/gpus), read Oct 24, 2026. [Qwen3-Embedding-4B on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/qwen3-embedding-4b/gpus) ## SAVRN's Notes on Qwen3-Embedding-4B Qwen3 Embedding 4B turns text into vectors for search and retrieval, built on the Qwen3 language model and supporting many languages with a 40,960-token window. It sits between the 0.6B and 8B models in the same series and draws about 2.1 million downloads a month. It needs about 10 GB at 16-bit. Apache 2.0 allows commercial use. For retrieval quality close to the top of the series at half the cost of the 8B model, this is the size we would test first. SAVRN Research, 2026-09-18 ## Model Card By Qwen, published under apache-2.0, revision 5cf2132abc99. ## Highlights The Qwen3 Embedding model series is the latest proprietary model of the Qwen family, specifically designed for text embedding and ranking tasks. Building upon the dense foundational models of the Qwen3 series, it provides a comprehensive range of text embeddings and reranking models in various sizes (0.6B, 4B, and 8B). This series inherits the exceptional multilingual capabilities, long-text understanding, and reasoning skills of its foundational model. The Qwen3 Embedding series represents significant advancements in multiple text embedding and ranking tasks, including text retrieval, code retrieval, text classification, text clustering, and bitext mining. **Exceptional Versatility**: The embedding model has achieved state-of-the-art performance across a wide range of downstream application evaluations. The 8B size embedding model ranks **No.1** in the MTEB multilingual leaderboard (as of June 5, 2025, score **70.58**), while the reranking model excels in various text retrieval scenarios. **Comprehensive Flexibility**: The Qwen3 Embedding series offers a full spectrum of sizes (from 0.6B to 8B) for both embedding and… [Read the full model card (2,148 words)](https://savrn.com/models/qwen3-embedding-4b/card) ## Configuration Architecture Qwen3ForCausalLM Context length (tokens) 40,960 Layers 36 Hidden size 2,560 Feed-forward size 9,728 Attention heads 32 Key/value heads 8 Head dimension 128 Vocabulary size 151,665 RoPE base 1,000,000 Stored precision bfloat16 Model type qwen3 ## Identity and Version Repository Qwen/Qwen3-Embedding-4B Publisher Qwen Task Feature extraction Modality Text Library sentence-transformers Parameters 4B parameters Languages Not stated by the source Revision 5cf2132abc99cad020ac570b19d031efec650f2b First published 2025-06-03 Last updated 2025-06-20 ## Files and Weights 14 files, 8.1 GB in total. The weights are 2 files totalling 8.0 GB in safetensors. Weights2 files · 8.0 GB Configuration6 files · 32.2 KB Tokenizer4 files · 15.9 MB Documentation1 file · 17.3 KB Repository1 file · 1.6 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | model-00001-of-00002.safetensors | Weights | 5.0 GB | e70bfe3c9705 | | model-00002-of-00002.safetensors | Weights | 3.1 GB | ed1b87c8e9eb | | 1_Pooling/config.json | Configuration | 313 B | — | | config.json | Configuration | 727 B | — | | config_sentence_transformers.json | Configuration | 215 B | — | | generation_config.json | Configuration | 117 B | — | | model.safetensors.index.json | Configuration | 30.4 KB | — | | modules.json | Configuration | 349 B | — | | README.md | Documentation | 17.3 KB | — | | .gitattributes | Repository | 1.6 KB | — | | merges.txt | Tokenizer | 1.7 MB | — | | tokenizer.json | Tokenizer | 11.4 MB | 83cdf8c3a34f | | tokenizer_config.json | Tokenizer | 7.3 KB | — | | vocab.json | Tokenizer | 2.8 MB | — | ## License and Download License apache-2.0 Access Open weights, no gate Download size 8.0 GB [Download from Qwen](https://modelscope.cn/models/Qwen/Qwen3-Embedding-4B) Released by Qwen through ModelScope. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Built From - Derived from [Qwen/Qwen3-4B-Base](https://savrn.com/models/qwen3-4b-base) - Described by [arXiv:2506.05176](https://savrn.com/papers/qwen3-embedding-advancing-text-embedding-and-reranking-through-foundation-models) ## Memory Requirements | Precision | Weights in memory | | --- | --- | | As published | 8.0 GB | | 16-bit | 8.0 GB | | 8-bit | 4.0 GB | | 4-bit | 2.0 GB | Weights only, from the published parameter count; the key-value cache and runtime add to this. ## Built on This Model - Quantized from[Qwen3-Embedding-4B-W4A16-G128](https://savrn.com/models/qwen3-embedding-4b-w4a16-g128) - Derived from[Qwen3-Embedding-4B-W4A16-G128](https://savrn.com/models/qwen3-embedding-4b-w4a16-g128) - Quantized from[RK182X-EMBEDDING-Qwen3-Embedding-4B](https://savrn.com/models/rk182x-embedding-qwen3-embedding-4b) - Derived from[RK182X-EMBEDDING-Qwen3-Embedding-4B](https://savrn.com/models/rk182x-embedding-qwen3-embedding-4b) - Quantized from[AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4](https://savrn.com/models/ax-qwen3-embedding-4b-cuda-axq-nvfp4-w4a4) - Derived from[AX-Qwen3-Embedding-4B-CUDA-AXQ-NVFP4-W4A4](https://savrn.com/models/ax-qwen3-embedding-4b-cuda-axq-nvfp4-w4a4) ## Compare Qwen3-Embedding-4B - [Qwen3-Embedding-4B vs Qwen3-Embedding-8B](https://savrn.com/models/compare/qwen3-embedding-4b-vs-qwen3-embedding-8b) ## Questions About Qwen3-Embedding-4B ### How much GPU memory does Qwen3-Embedding-4B need? About 9.7 GB at 16-bit and 2.4 GB at 4-bit: the weights (4B parameters) plus a working margin. A long context needs more. ### What is the cheapest GPU to run Qwen3-Embedding-4B on? At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists. ### Can I use Qwen3-Embedding-4B commercially? Yes. Qwen3-Embedding-4B is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors. ### What is Qwen3-Embedding-4B's context length? 40,960 tokens, from the maximum position embeddings in its published configuration. ## Similar Models Model · Feature extraction ### [social-jev-4B-preview](https://savrn.com/models/social-jev-4b-preview) [Social Jev](https://savrn.com/model-publishers/social-jev) SocialJev is a family of classification foundation models for the social sciences, inspired by TypeSafe’s Jev. The models combine the flexibility of modern language models pretrained on trillions of tokens of text with the speed of lightweight NLP methods. Define your task through a simple typed interface and SocialJev returns typed outputs. Note: SocialJev is under active development. Interfaces may change between early releases. The model can be used directly through Transformers without installing the SocialJev package: Results are ordered first by document, then by query for that document. Each result provides.choice and.probabilities; the probabilities above are illustrative. The… Open weights 4B parameters transformers [View model](https://savrn.com/models/social-jev-4b-preview) Model · Feature extraction ### [Qwen3-Embedding-4B-W4A16-G128](https://savrn.com/models/qwen3-embedding-4b-w4a16-g128) [Mou Geren](https://savrn.com/model-publishers/boboliu) GPTQ Quantized Qwen/Qwen3-Embedding-4B with THUIR/T2Ranking and m-a-p/COIG-CQIA for calibration set. ~0.72% lost in C-MTEB. Evaluation performed with official code. pip install compressed-tensors optimum and auto-gptq / gptqmodel, then goto the official usage guide. Open weights apache-2.0 4.1B parameters 40,960 tokens sentence-transformers [View model](https://savrn.com/models/qwen3-embedding-4b-w4a16-g128) Model · Feature extraction ### [MicroGlot](https://savrn.com/models/microglot) [Athan Zhaohong Li](https://savrn.com/model-publishers/athanzli) A taxonomy-informed sparse DNA foundation model for microbial genomics. MicroGlot is a 23-layer decoder-only mixture-of-experts transformer pretrained on 378.3 billion nucleotides from 3.70 million sequences across 99,700 microbial species, spanning bacteria, archaea, fungi, protists, viruses and plasmids. It encodes the taxonomic hierarchy as hyperbolic (Poincaré) embeddings and uses them both as an input token and to steer expert routing. Paper: A Taxonomy-Informed Sparse DNA Foundation Model for Microbial Genomics\ A species is known if it is one of the 99,700 pretraining species (check with tokenizer.hasspecies(name)). For taxonomic classification tasks, use MicroGlot-plain, as… Open weights cc-by-4.0 3B parameters transformers [View model](https://savrn.com/models/microglot) Model · Feature extraction ### [kb-wemm-embedding-2b-mlx-mixed](https://savrn.com/models/kb-wemm-embedding-2b-mlx-mixed) [Guodong Fei](https://savrn.com/model-publishers/kn90688) Native MLX conversion of ewin-reg/WeMM-Embedding-2B-Quantized — the hybrid FP8-protected attention + group-16 INT4 checkpoint — that preserves its mixed-precision allocation instead of flattening everything to one precision. The source checkpoint is curvature-guided: it deliberately keeps the precision-sensitive path at ~8-bit (FP8 E4M3) and compresses only the quantization-robust masses to INT4. Uniform 4-bit quantization destroys that — the source README itself documents softmax attention collapse (DKL >= 7.50) when Q/K go to 4 bits. This build mirrors the Two deviations from the source, both improvements or forced by MLX: - 8-bit group-64 is finer than the source's per-tensor-scalar FP8… Open weights other 2.2B parameters 262,144 tokens mlx [View model](https://savrn.com/models/kb-wemm-embedding-2b-mlx-mixed) Model · Feature extraction ### [MOSS-Audio-Tokenizer-v2](https://savrn.com/models/moss-audio-tokenizer-v2) [Le Jin](https://savrn.com/model-publishers/jinstudio) Open weights apache-2.0 2.1B parameters transformers [View model](https://savrn.com/models/moss-audio-tokenizer-v2) Model · Feature extraction ### [tiny-audio-granite-qwen](https://savrn.com/models/tiny-audio-granite-qwen) [Alex Kroman](https://savrn.com/model-publishers/mazesmazes) This is the model card of a transformers model that has been pushed on the Hub. Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019). Open weights 1.9B parameters transformers [View model](https://savrn.com/models/tiny-audio-granite-qwen) ## Qwen [All models and datasets](https://savrn.com/model-publishers/qwen) ## Versions - [5cf2132abc99](https://savrn.com/models/qwen3-embedding-4b/versions/5cf2132abc99) · current 2026-10-24 ## Explore More - [All feature extraction models](https://savrn.com/models/tasks/feature-extraction) - [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0) - [Model comparisons](https://savrn.com/models/comparisons) - [The model directory](https://savrn.com/models) - [Open model prices by host](https://savrn.com/ai-index/pricing/open-models) ## Source - Repository metadata, read 2026-10-19. - [Qwen website](https://qwen.ai/) - [Hugging Face record](https://huggingface.co/Qwen/Qwen3-Embedding-4B) - [How the hub is built](https://savrn.com/model-hub/methodology)