Aether 2.5 Pro is a fine-tuned version of Qwen2.5-3B-Instruct (SFT + LoRA). It offers improved reasoning, stronger instruction following, and better multilingual performance in German and English.

🖥️ Easiest way to run the model:
Download MonoAIStudio – our local chat application.
It comes pre-installed with Aether 2.5, Aether 2.5 Pro and Aether 2.5 Coder.
👉 Download MonoAIStudio.zip

🔗 Looking for the LoRA Adapter?
👉 Maxilicious20/Aether-2.5-Pro

📦 Available Quantizations

Filename	Quantization	Quality	Size	Recommendation
`aether-2.5-pro-f16.gguf`	F16	Maximum	~6.0 GB	Best quality
`aether-2.5-pro-q8_0.gguf`	Q8_0	Very High	~3.2 GB	Excellent quality
`aether-2.5-pro-q4_k_m.gguf`	Q4_K_M	Balanced	~1.9 GB	Recommended

🚀 How to Run

1. MonoAIStudio (Recommended)

Download MonoAIStudio.zip
Extract it
Run MonoAIStudio.exe
All Aether 2.5 models are already included

2. LM Studio

Open LM Studio
Search for Maxilicious20/Aether-2.5-Pro-GGUF
Download aether-2.5-pro-q4_k_m.gguf
Load the model

3. llama.cpp

./llama-cli -m aether-2.5-pro-q4_k_m.gguf -p "Hello!" -n 256

Maxilicious20/Aether-2.5-Pro-GGUF

Aether 2.5 Pro - GGUF

📦 Available Quantizations

🚀 How to Run

1. MonoAIStudio (Recommended)

2. LM Studio

3. llama.cpp

Maxilicious20/
Aether-2.5-Pro-GGUF