granite-embedding-small-english-r2 vs all-MiniLM-L6-v2
Specs, VRAM requirements and download trends — updated 10 November 2026.
ibm-granite · A · Open weights
Params
50 M
Context
8,192
30d
6.9 M
apache-2.0 · commercial OK
Xenova · B · Open weights
Params
20 M
Context
512
30d
4.5 M
apache-2.0 · commercial OK
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | granite-embedding-small-english-r2 | all-MiniLM-L6-v2 |
|---|---|---|
| Parameters | 50 M | 20 M |
| Architecture | ModernBertModel | BertModel |
| Context length | 8,192 | 512 |
| Licence | apache-2.0 | apache-2.0 |
| Commercial use | Allowed | Allowed |
| Languages | 1 | — |
| Downloads 30d | 6,856,446 | 4,506,671 |
| Downloads all time | 34.9 M | 35.7 M |
| Quantizations on the Hub | SAFETENSORS | — |
| Gated | No | No |
| First seen on the Hub | 2025-07-17 | 2023-05-02 |
Download trend
Daily snapshots, last 90 days (13 Aug – 10 Nov)
6.9 M2.5 M
granite-embedding-small-english-r2
all-MiniLM-L6-v2
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
granite-embedding-small-english-r2 · fp16
0.6 GB / 24 GB