bge-m3 vs paraphrase-multilingual-mpnet-base-v2
Specs, VRAM requirements and download trends — updated 10 November 2026.
sentence-transformers · B · Open weights
Params
280 M
Context
512
30d
9.1 M
apache-2.0 · commercial OK
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | bge-m3 | paraphrase-multilingual-mpnet-base-v2 |
|---|---|---|
| Parameters | — | 280 M |
| Architecture | XLMRobertaModel | XLMRobertaModel |
| Context length | 8,192 | 512 |
| Licence | mit | apache-2.0 |
| Commercial use | Allowed | Allowed |
| Languages | — | 50 |
| Downloads 30d | 29,564,329 | 9,113,974 |
| Downloads all time | 362.4 M | 149.4 M |
| Quantizations on the Hub | — | SAFETENSORS |
| Gated | No | No |
| First seen on the Hub | 2024-01-27 | 2022-03-02 |
Download trend
Daily snapshots, last 90 days (13 Aug – 10 Nov)
38.2 M8.6 M
bge-m3
paraphrase-multilingual-mpnet-base-v2
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
paraphrase-multilingual-mpnet-base-v2 · fp16
1.8 GB / 24 GB