bartowski
Qwen2.5 3B Instruct
Qwen2.5 3B Instruct is a 3.09B dense qwen2 model from bartowski. Configured context length is 32,768 tokens. License: other. Quant sizes are the GGUF file bytes published on Hugging Face when a file exists.
Parameters3.1B
Context32K
Attentiongqa
Licenseother
36 layers · 2 KV heads · head dim 128 · GGUF repo · config
Quant file sizes
| Quant | File size | Source |
|---|---|---|
| F16 | 5.75 GB | Hugging Face file bytes |
| Q8_0 | 3.06 GB | Hugging Face file bytes |
| Q6_K | 4.80 GB | Hugging Face file bytes |
| Q5_K_M | 2.07 GB | Hugging Face file bytes |
| Q5_K_S | 2.02 GB | Hugging Face file bytes |
| Q4_K_M | 1.80 GB | Hugging Face file bytes |
| Q4_K_S | 1.71 GB | Hugging Face file bytes |
| IQ4_XS | 1.62 GB | Hugging Face file bytes |
| Q4_0 | 6.80 GB | Hugging Face file bytes |
| Q3_K_L | 1.59 GB | Hugging Face file bytes |
| Q3_K_M | 1.48 GB | Hugging Face file bytes |
| IQ3_M | 1.39 GB | Hugging Face file bytes |
| Q3_K_S | 1.35 GB | Hugging Face file bytes |
| Q2_K | 2.44 GB | Hugging Face file bytes |
On common machines
| Device | Verdict | Quant | Speed |
|---|---|---|---|
| GeForce RTX 4060 Ti 8GB | Tight fit | F16 | 24–33 |
| GeForce RTX 3060 12GB | Runs great | F16 | 31–42 |
| GeForce RTX 4060 Ti 16GB | Runs great | F16 | 24–33 |
| GeForce RTX 4090 | Runs great | F16 | 86–117 |
| GeForce RTX 5090 | Runs great | F16 | 152–207 |
| Apple M4 Pro | Runs great | F16 | 23–32 |
| Apple M5 | Runs great | F16 | 13–18 |
| CPU only · DDR5-5600 | Runs on CPU | F16 | 7.6–10 |
Smallest devices that fit it fully
- GeForce RTX 3050 6GB — Runs, Q8_0, 26–35 tok/s
- GeForce RTX 2060 6GB — Runs, Q8_0, 52–70 tok/s
- GeForce GTX 1660 Super — Runs, Q8_0, 52–70 tok/s
- GeForce GTX 1660 Ti — Runs, Q8_0, 44–60 tok/s
- Intel Arc A380 — Runs, Q8_0, 29–39 tok/s
- GeForce RTX 5060 Ti 8GB — Tight fit, F16, 38–52 tok/s
Run it
Commands for the GeForce RTX 4060 Ti 16GB estimate (F16).
llama.cpp
llama-server -hf bartowski/Qwen2.5-3B-Instruct-GGUF:F16 -c 8192 -ngl 36
Ollama
ollama run qwen2.5
LM Studio
lms get bartowski/Qwen2.5-3B-Instruct-GGUF # Then load quant F16