Nemotron 3.5 Bolt (E4B, Q4)
NVIDIA's conversational model built on Gemma 4 E4B, with native tool calling, optional reasoning, and a 64K context window. Text only. The official Q4_K_M download is about 6.2GB; working memory depends on context and hardware.
| At a glance | |
|---|---|
| Made by | NVIDIA |
| Download size | 6.2 GB |
| Parameters | 8B |
| Quantization | Q4_K_M |
| Context window | 66K tokens |
| Uses tools | Yes |
| Best for | Everyday work |
| Runs on | llama.cpp |
| License | OpenMDW 1.1 |
| Added to the catalog | 2026-10-25 |
Gezel downloads this model the first time a gezel needs it, checks the file against its published hash, and applies tuned settings for it.