unsloth
Qwen3 14B
Qwen3 14B is a 14.8B dense qwen3 model from unsloth. Configured context length is 40,960 tokens. License: apache-2.0. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters14.8B
Context40K
Attentiongqa
Licenseapache-2.0
40 layers · 8 KV heads · head dim 128 · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| BF16 | 27.5 GB | Hugging Face file bytes |
| UD-Q8_K_XL | 17.5 GB | Hugging Face file bytes |
| Q8_0 | 14.6 GB | Hugging Face file bytes |
| UD-Q6_K_XL | 12.4 GB | Hugging Face file bytes |
| Q6_K | 11.3 GB | Hugging Face file bytes |
| UD-Q5_K_XL | 9.82 GB | Hugging Face file bytes |
| Q5_K_M | 9.79 GB | Hugging Face file bytes |
| Q5_K_S | 9.56 GB | Hugging Face file bytes |
| UD-Q4_K_XL | 8.53 GB | Hugging Face file bytes |
| Q4_K_M | 8.38 GB | Hugging Face file bytes |
| Q4_K_S | 7.98 GB | Hugging Face file bytes |
| IQ4_NL | 7.95 GB | Hugging Face file bytes |
| IQ4_XS | 7.58 GB | Hugging Face file bytes |
| Q4_1 | 8.74 GB | Hugging Face file bytes |
| Q4_0 | 7.96 GB | Hugging Face file bytes |
| UD-Q3_K_XL | 7.07 GB | Hugging Face file bytes |
| Q3_K_M | 6.82 GB | Hugging Face file bytes |
| Q3_K_S | 6.20 GB | Hugging Face file bytes |
| UD-IQ3_XXS | 5.60 GB | Hugging Face file bytes |
| UD-Q2_K_XL | 5.65 GB | Hugging Face file bytes |
| Q2_K | 10.9 GB | Hugging Face file bytes |
| UD-IQ2_XXS | 4.16 GB | Hugging Face file bytes |
| UD-IQ1_M | 3.79 GB | Hugging Face file bytes |
| UD-IQ1_S | 3.56 GB | Hugging Face file bytes |
On common machines
| Device | Verdict | Quant | Speed |
|---|---|---|---|
| GeForce RTX 4060 Ti 8GB | Tight fit | UD-IQ3_XXS | 24–32 |
| GeForce RTX 3060 12GB | Tight fit | Q5_K_S | 18–25 |
| GeForce RTX 4060 Ti 16GB | Tight fit | UD-Q6_K_XL | 11–15 |
| GeForce RTX 4090 | Tight fit | UD-Q8_K_XL | 28–38 |
| GeForce RTX 5090 | Tight fit | BF16 | 33–44 |
| Apple M4 Pro | Tight fit | Q8_0 | 8.8–12 |
| Apple M5 | Tight fit | Q5_K_S | 7.7–10 |
| CPU only · DDR5-5600 | Runs on CPU | UD-Q8_K_XL | 2.5–3.3 |
Smallest devices that fit it fully
- GeForce RTX 3050 6GB — Tight fit, UD-IQ1_M, 20–27 tok/s
- GeForce RTX 2060 6GB — Tight fit, UD-IQ1_M, 39–53 tok/s
- GeForce GTX 1660 Super — Tight fit, UD-IQ1_M, 39–53 tok/s
- GeForce GTX 1660 Ti — Tight fit, UD-IQ1_M, 33–46 tok/s
- Intel Arc A380 — Tight fit, UD-IQ1_M, 22–29 tok/s
- GeForce RTX 5060 Ti 8GB — Tight fit, UD-IQ3_XXS, 37–50 tok/s
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (UD-Q6_K_XL).
llama.cpp
llama-server -hf unsloth/Qwen3-14B-GGUF:UD-Q6_K_XL -c 8192 -ngl 40
Ollama
ollama run qwen3
LM Studio
lms get unsloth/Qwen3-14B-GGUF # Then load quant UD-Q6_K_XL