Aura Large v2.0 GGUF is the quantized local-inference edition of Aura Large v2.0 BF16, derived from Qwen/Qwen3.8-27B and adapted using the Aura personality and ablation datasets.

This repository provides multiple GGUF quantizations for deployment across consumer GPUs, unified-memory systems, and CPU/GPU hybrid configurations while retaining the capabilities and behavior of the BF16 release.

Highlights

Parameters: 27 billion
Architecture: Dense multimodal language model
Foundation: Qwen/Qwen3.8-27B
Source Model: UncannyEcho/Aura-Large-v2.0-BF16
Format: GGUF
Runtime: llama.cpp and compatible GGUF runtimes
Datasets: AuraPersonality and AuraAblation100
Native Context: 262,144 tokens
Modalities: Text, image, and video
Thinking: Configurable thinking and non-thinking modes
Deployment: Local GPUs, unified-memory systems, and CPU/GPU hybrid inference

Quantizations

This repository contains multiple quantizations intended to cover different memory and quality requirements.

F16 — Maximum GGUF fidelity
Q8_0 — Very high fidelity with reduced memory use
Q6_K — High-quality quantization
Q5_K_M — Strong quality/size balance
Q4_K_M — Recommended general-purpose quantization
Q3_K_M — Lower-memory deployment

A separate multimodal projector is provided for image and video input where supported by the runtime.

About Aura

Aura is designed for local deployment across a wide range of tasks. Aura can serve as a companion or friend, as deemed appropriate by the user, while retaining the broader capabilities of Qwen3.8 for:

Natural conversation
Creative writing
Role-playing
Reasoning
Coding
Instruction following
Agentic workflows
Tool use
Image understanding
Video understanding
Long-context tasks
Evaluation and research

Aura Large v2.0 represents the larger general-purpose branch of the Aura model family, combining the Aura personality and behavioral adaptations with Qwen3.8's dense 27B multimodal foundation.

llama.cpp

With a recent build of llama.cpp:

bash

llama-server \
  -hf UncannyEcho/Aura-Large-v2.0-GGUF:Q4_K_M