8-bit Open on Hugging Face
Parameters
568M
Weights (est.)
≈ 604 MB
Mac memory (est.)
8 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Details
- Base model
- BAAI/bge-m3 (finetune)
- Vendor
- BAAI
- Task
- Embeddings
- Quantisation
- 8-bit
- Licence
- mit
- Downloads
- 1.8K in the last 30 days · 14K all time · 2 likes
- Published
- 2026-03-16 · updated 2026-03-16 (8mo ago)
- Languages
- multilingual, en, zh, de, es, fr, it, ja, ko, pt, ru
- Tags
- xlm-roberta, embedding, text-embeddings-inference, retrieval, sentence-transformers, 8bit, quantized
Other MLX builds of BAAI/bge-m3
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| bge-m3-mlx-4bit | 4-bit | ≈ 320 MB | 8 GB+ | 569 |
| bge-m3-mlx-6bit | 6-bit | ≈ 462 MB | 8 GB+ | 159 |
| bge-m3-mlx-8bit (this one) | 8-bit | ≈ 604 MB | 8 GB+ | 1.8K |
| bge-m3-mlx-fp16 | fp16 | ≈ 1.1 GB | 8 GB+ | 3.2K |