Gemma 4 (E2B, Q4)
Google's edge-optimized Gemma 4 at ~2B effective parameters. Multimodal (text + image + audio), native tool calling. Quantization-aware-trained 4-bit weights — the safe small-model default for any modern laptop.
| At a glance | |
|---|---|
| Made by | |
| Download size | 2.6 GB |
| Parameters | 2.3B |
| Quantization | UD-Q4_K_XL |
| Context window | 128K tokens |
| Uses tools | Yes |
| Best for | Everyday work |
| Runs on | llama.cpp, MLX, Ollama |
| License | Apache 2.0 |
| Added to the catalog | 2026-08-28 |
Gezel downloads this model the first time a gezel needs it, checks the file against its published hash, and applies tuned settings for it.