unsloth
gemma 3 27b it
gemma 3 27b it is a 27.0B dense gemma3 model from unsloth. Configured context length is 131,072 tokens. License: gemma. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters27.0B
Context128K
Attentionhybrid
Licensegemma
62 layers · 16 KV heads · head dim 128 · 52 sliding layers · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| BF16 | 50.3 GB | Hugging Face file bytes |
| UD-Q8_K_XL | 29.6 GB | Hugging Face file bytes |
| Q8_0 | 26.7 GB | Hugging Face file bytes |
| UD-Q6_K_XL | 22.1 GB | Hugging Face file bytes |
| Q6_K | 20.6 GB | Hugging Face file bytes |
| UD-Q5_K_XL | 18.0 GB | Hugging Face file bytes |
| Q5_K_M | 17.9 GB | Hugging Face file bytes |
| Q5_K_S | 17.5 GB | Hugging Face file bytes |
| UD-Q4_K_XL | 15.7 GB | Hugging Face file bytes |
| Q4_K_M | 15.4 GB | Hugging Face file bytes |
| Q4_K_S | 14.6 GB | Hugging Face file bytes |
| IQ4_NL | 14.5 GB | Hugging Face file bytes |
| IQ4_XS | 13.8 GB | Hugging Face file bytes |
| Q4_1 | 16.0 GB | Hugging Face file bytes |
| Q4_0 | 14.5 GB | Hugging Face file bytes |
| UD-Q3_K_XL | 12.8 GB | Hugging Face file bytes |
| Q3_K_M | 12.5 GB | Hugging Face file bytes |
| Q3_K_S | 11.3 GB | Hugging Face file bytes |
| UD-IQ3_XXS | 10.1 GB | Hugging Face file bytes |
| UD-Q2_K_XL | 9.95 GB | Hugging Face file bytes |
| Q2_K | 19.6 GB | Hugging Face file bytes |
| UD-IQ2_XXS | 7.31 GB | Hugging Face file bytes |
| UD-IQ1_M | 6.51 GB | Hugging Face file bytes |
| UD-IQ1_S | 6.06 GB | Hugging Face file bytes |
On common machines
| Device | Verdict | Quant | Speed |
|---|---|---|---|
| GeForce RTX 4060 Ti 8GB | Tight fit | UD-IQ1_S | 23–31 |
| GeForce RTX 3060 12GB | Runs | UD-IQ2_XXS | 16–22 |
| GeForce RTX 4060 Ti 16GB | Tight fit | UD-Q3_K_XL | 8.9–12 |
| GeForce RTX 4090 | Tight fit | Q6_K | 23–31 |
| GeForce RTX 5090 | Tight fit | Q8_0 | 30–41 |
| Apple M4 Pro | Tight fit | UD-Q4_K_XL | 7.2–9.8 |
| Apple M5 | Runs | UD-IQ2_XXS | 7.0–9.6 |
| CPU only · DDR5-5600 | Runs on CPU | UD-Q6_K_XL | 1.8–2.4 |
Smallest devices that fit it fully
- GeForce RTX 5060 Ti 8GB — Tight fit, UD-IQ1_S, 35–48 tok/s
- GeForce RTX 5060 — Tight fit, UD-IQ1_S, 35–48 tok/s
- GeForce RTX 5050 — Tight fit, UD-IQ1_S, 25–34 tok/s
- GeForce RTX 4060 Ti 8GB — Tight fit, UD-IQ1_S, 23–31 tok/s
- GeForce RTX 4060 — Tight fit, UD-IQ1_S, 21–29 tok/s
- GeForce RTX 3070 Ti — Tight fit, UD-IQ1_S, 48–65 tok/s
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (UD-Q3_K_XL).
llama.cpp
llama-server -hf unsloth/gemma-3-27b-it-GGUF:UD-Q3_K_XL -c 8192 -ngl 62
Ollama
ollama run gemma
LM Studio
lms get unsloth/gemma-3-27b-it-GGUF # Then load quant UD-Q3_K_XL