multilingual-e5-base vs embeddinggemma-300m
Specs, VRAM requirements and download trends — updated 21 September 2026.
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | multilingual-e5-base | embeddinggemma-300m |
|---|---|---|
| Parameters | 280 M | 300 M |
| Architecture | XLMRobertaModel | Gemma3TextModel |
| Context length | 512 | — |
| Licence | mit | gemma |
| Commercial use | Allowed | Conditional |
| Languages | 94 | — |
| Downloads 30d | 7,570,279 | 2,674,650 |
| Downloads all time | 65.4 M | 18.6 M |
| Quantizations on the Hub | SAFETENSORS | SAFETENSORS |
| Gated | No | Yes — accept licence |
| First seen on the Hub | 2023-05-19 | 2025-07-17 |
Download trend
Daily snapshots, last 56 days (28 Jul – 21 Sep)
7.6 M1.9 M
multilingual-e5-base
embeddinggemma-300m
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
multilingual-e5-base · fp16
1.8 GB / 24 GB
embeddinggemma-300m · fp16
1.9 GB / 24 GB