unsloth
gemma 4 E4B it
gemma 4 E4B it is a 7.52B dense gemma4 model from unsloth. Configured context length is 131,072 tokens. License: apache-2.0. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters7.5B
Context128K
Attentionhybrid
Licenseapache-2.0
42 layers · 2 KV heads · head dim 256 · 35 sliding layers · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| BF16 | 14.0 GB | Hugging Face file bytes |
| UD-Q8_K_XL | 8.11 GB | Hugging Face file bytes |
| Q8_0 | 7.63 GB | Hugging Face file bytes |
| UD-Q6_K_XL | 6.95 GB | Hugging Face file bytes |
| Q6_K | 6.59 GB | Hugging Face file bytes |
| UD-Q5_K_XL | 6.20 GB | Hugging Face file bytes |
| Q5_K_M | 5.11 GB | Hugging Face file bytes |
| Q5_K_S | 5.03 GB | Hugging Face file bytes |
| UD-Q4_K_XL | 4.77 GB | Hugging Face file bytes |
| Q4_K_M | 4.64 GB | Hugging Face file bytes |
| Q4_K_S | 4.51 GB | Hugging Face file bytes |
| IQ4_NL | 4.50 GB | Hugging Face file bytes |
| IQ4_XS | 4.39 GB | Hugging Face file bytes |
| Q4_1 | 4.73 GB | Hugging Face file bytes |
| Q4_0 | 4.50 GB | Hugging Face file bytes |
| UD-Q3_K_XL | 4.27 GB | Hugging Face file bytes |
| Q3_K_M | 3.78 GB | Hugging Face file bytes |
| Q3_K_S | 3.60 GB | Hugging Face file bytes |
| UD-IQ3_XXS | 3.46 GB | Hugging Face file bytes |
| UD-Q2_K_XL | 3.50 GB | Hugging Face file bytes |
On common machines
| Device | Verdict | Quant | Speed |
|---|---|---|---|
| GeForce RTX 4060 Ti 8GB | Tight fit | UD-Q5_K_XL | 22–29 |
| GeForce RTX 3060 12GB | Runs | UD-Q8_K_XL | 21–29 |
| GeForce RTX 4060 Ti 16GB | Runs great | UD-Q8_K_XL | 17–23 |
| GeForce RTX 4090 | Runs | BF16 | 35–48 |
| GeForce RTX 5090 | Runs great | BF16 | 63–85 |
| Apple M4 Pro | Runs | BF16 | 9.5–13 |
| Apple M5 | Runs | UD-Q8_K_XL | 8.9–12 |
| CPU only · DDR5-5600 | Runs on CPU | BF16 | 3.1–4.3 |
Smallest devices that fit it fully
- GeForce RTX 3050 6GB — Tight fit, Q4_K_M, 17–24 tok/s
- GeForce RTX 2060 6GB — Tight fit, Q4_K_M, 35–47 tok/s
- GeForce GTX 1660 Super — Tight fit, Q4_K_M, 35–47 tok/s
- GeForce GTX 1660 Ti — Tight fit, Q4_K_M, 30–41 tok/s
- Intel Arc A380 — Tight fit, Q4_K_M, 19–26 tok/s
- GeForce RTX 5060 Ti 8GB — Tight fit, UD-Q5_K_XL, 33–46 tok/s
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (UD-Q8_K_XL).
llama.cpp
llama-server -hf unsloth/gemma-4-E4B-it-GGUF:UD-Q8_K_XL -c 8192 -ngl 42
Ollama
ollama run gemma
LM Studio
lms get unsloth/gemma-4-E4B-it-GGUF # Then load quant UD-Q8_K_XL