jina embeddings v4 text retrieval needs about 2.7 GB of memory at Q4_K_M (a 1.9 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 18 quants, from 0.8 GB (IQ1_S) to 6.2 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 1.2 GB.
Jina Embeddings v4 text retrieval – multimodal, multilingual embeddings
Jina Embeddings v4 text retrieval is a 3.1B parameter model built on Qwen2.5-VL-3B-Instruct. It produces dense embeddings for text, images, and visual documents, supporting retrieval, text matching, and code tasks. The model is available in multiple GGUF quantizations for llama.cpp.
Good at
- Dense embeddings for multimodal retrieval
- Multilingual support for 30+ languages
- Task‑specific adapters for retrieval, text‑matching, and code
Limits
- No vision support in text‑only GGUF variants
- Cannot output multi‑vector embeddings in text‑only variants
- License terms not specified
Best for
- Embedding visual documents with charts and tables
- Retrieval in multilingual corpora
- Code snippet similarity search
Who it’s for. Developers building retrieval systems that need both text and image embeddings. Researchers working on multimodal document search and multilingual retrieval tasks.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 0.8 GB | 1.6 GB | Fits | Get file → | |
| IQ1_M | 0.9 GB | 1.7 GB | Fits | Get file → | |
| IQ2_XXS | 0.9 GB | 1.8 GB | Fits | Get file → | |
| IQ2_M | 1.1 GB | 1.9 GB | Fits | Get file → | |
| Q2_K | 1.3 GB | 2.1 GB | Fits | Get file → | |
| IQ3_XXS | 1.3 GB | 2.1 GB | Fits | Get file → | |
| IQ3_XS | 1.4 GB | 2.2 GB | Fits | Get file → | |
| IQ3_S | 1.5 GB | 2.3 GB | Fits | Get file → | |
| IQ3_M | 1.5 GB | 2.3 GB | Fits | Get file → | |
| Q3_K_M | 1.6 GB | 2.4 GB | Fits | Get file → | |
| IQ4_XS | 1.7 GB | 2.5 GB | Fits | Get file → | |
| IQ4_NL | 1.8 GB | 2.6 GB | Fits | Get file → | |
| Q4_K_M | 1.9 GB | 2.7 GB | Fits | Get file → | |
| Q5_K_S | 2.2 GB | 3.0 GB | Fits | Get file → | |
| Q5_K_M | 2.2 GB | 3.0 GB | Fits | Get file → | |
| Q6_K | 2.5 GB | 3.3 GB | Fits | Get file → | |
| Q8_0 | 3.3 GB | 4.1 GB | Fits | Get file → | |
| F16 | 6.2 GB | 7.0 GB | Fits | Get file → |
For 8 GB VRAM, use IQ1_S (0.79 GB, ~1.6 GB RAM) or IQ1_M (0.85 GB, ~1.7 GB RAM). If 12 GB VRAM is available, Q4_K_M (1.93 GB, ~2.7 GB RAM) offers higher precision. Larger models such as Q8_0 or F16 require 4–7 GB RAM and are suitable only on high‑end GPUs.
What GPU do you need for jina embeddings v4 text retrieval?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: the best quant that fits is Q6_K (2.5 GB file, about 3.3 GB in use).
- 6 GB VRAM: the best quant that fits is Q8_0 (3.3 GB file, about 4.1 GB in use).
- 8 GB VRAM: the best quant that fits is F16 (6.2 GB file, about 7.0 GB in use).
Which jina embeddings v4 text retrieval quant should you choose?
Q4_K_M is the usual starting point for jina embeddings v4 text retrieval: it balances file size and quality well. Here it is a 1.9 GB file.
The Q8_0 file is 3.3 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is IQ1_S at 0.8 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about jina embeddings v4 text retrieval
How much VRAM does jina embeddings v4 text retrieval need?
About 2.7 GB at Q4_K_M with an 8K context window. That is the 1.9 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run jina embeddings v4 text retrieval on an 8 GB GPU?
Yes, with F16: the 6.2 GB file needs about 7.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of jina embeddings v4 text retrieval GGUF?
It depends on the quant: IQ1_S is 0.8 GB, IQ1_M is 0.9 GB, IQ2_XXS is 0.9 GB, IQ2_M is 1.1 GB, Q2_K is 1.3 GB, IQ3_XXS is 1.3 GB and 12 more.
What context length does jina embeddings v4 text retrieval support?
The GGUF header lists a trained context length of 128,000 tokens (125K). The KV cache grows with the context you use: about 0.3 GB at 8K and 1.2 GB at 32K.
Where can I download jina embeddings v4 text retrieval GGUF?
The files are on Hugging Face in the jinaai/jina-embeddings-v4-text-retrieval-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
Can I run this model on 8 GB VRAM?
Yes, use IQ1_S or IQ1_M quantizations; they fit within 8 GB VRAM and need about 1.6–1.7 GB RAM.
Does the model support image input?
Only the multimodal GGUF variants support image input; text‑only variants do not accept images.
Can I get multi‑vector embeddings?
Multi‑vector embeddings are available only in the full multimodal model, not in the text‑only GGUF versions.