---
license: apache-2.0
language: [en]
tags: [embeddings, gguf, ggml, text-embeddings, bert, crispembed]
pipeline_tag: feature-extraction
base_model: mixedbread-ai/mxbai-rerank-base-v1
---

# mxbai-rerank-base-v1 GGUF

GGUF format of [mixedbread-ai/mxbai-rerank-base-v1](https://huggingface.co/mixedbread-ai/mxbai-rerank-base-v1) for use with [CrispEmbed](https://github.com/CrispStrobe/CrispEmbed).

MixedBread Rerank Base v1. Cross-encoder reranker for English, good quality/speed balance.

## Files

| File | Quantization | Size |
|------|-------------|------|
| [mxbai-rerank-base-v1-q4_k.gguf](https://huggingface.co/cstr/mxbai-rerank-base-v1-GGUF/resolve/main/mxbai-rerank-base-v1-q4_k.gguf) | Q4_K | 149 MB |
| [mxbai-rerank-base-v1-q8_0.gguf](https://huggingface.co/cstr/mxbai-rerank-base-v1-GGUF/resolve/main/mxbai-rerank-base-v1-q8_0.gguf) | Q8_0 | 190 MB |
| [mxbai-rerank-base-v1.gguf](https://huggingface.co/cstr/mxbai-rerank-base-v1-GGUF/resolve/main/mxbai-rerank-base-v1.gguf) | F32 | 706 MB |


## Quick Start

```bash
# Download
huggingface-cli download cstr/mxbai-rerank-base-v1-GGUF mxbai-rerank-base-v1-q4_k.gguf --local-dir .

# Run with CrispEmbed
./crispembed -m mxbai-rerank-base-v1-q4_k.gguf "Hello world"

# Or with auto-download
./crispembed -m mxbai-rerank-base-v1 "Hello world"
```

## Model Details

| Property | Value |
|----------|-------|
| Architecture | BERT |
| Parameters | 86M |
| Embedding Dimension | 768 |
| Layers | 12 |
| Pooling | CLS |
| Tokenizer | WordPiece |
| Base Model | [mixedbread-ai/mxbai-rerank-base-v1](https://huggingface.co/mixedbread-ai/mxbai-rerank-base-v1) |

## Verification

Verified bit-identical to HuggingFace sentence-transformers (cosine similarity >= 0.999 on test texts).

## Usage with CrispEmbed

CrispEmbed is a lightweight C/C++ text embedding inference engine using ggml.
No Python runtime, no ONNX. Supports BERT, XLM-R, Qwen3, and Gemma3 architectures.

```bash
# Build CrispEmbed
git clone https://github.com/CrispStrobe/CrispEmbed
cd CrispEmbed
cmake -S . -B build && cmake --build build -j

# Encode
./build/crispembed -m mxbai-rerank-base-v1-q4_k.gguf "query text"

# Server mode
./build/crispembed-server -m mxbai-rerank-base-v1-q4_k.gguf --port 8080
curl -X POST http://localhost:8080/v1/embeddings \
    -d '{"input": ["Hello world"], "model": "mxbai-rerank-base-v1"}'
```

## Credits

- Original model: [mixedbread-ai/mxbai-rerank-base-v1](https://huggingface.co/mixedbread-ai/mxbai-rerank-base-v1)
- Inference engine: [CrispEmbed](https://github.com/CrispStrobe/CrispEmbed) (ggml-based)
- Conversion: `convert-bert-embed-to-gguf.py`