A small, fully local model that reasons and answers like a frontier teacher. Built by distilling glm-5.2 general-knowledge responses into a Qwen3-8B student with a QLoRA fine-tune, then merged, converted, and quantized to GGUF. The student base retains native tool calling and a long context window.

Capabilities

Size	Context	Input	Output	Tools
5.0 GB	128K	🅣 Text	Text	✅


Base model	Qwen/Qwen3-8B
Teacher	glm-5.2 (responses distilled, thinking disabled)
Method	4-bit QLoRA SFT (rank 16) → merge → GGUF
Quantization	Q4_K_M (~5 GB)
License	Apache-2.0 (inherited from Qwen3-8B)
Context	128K (inherited from base)
Tool calling	Supported (inherited from base)

Run it

Pull directly into Ollama:

ollama run hf.co/AdvancedDataIntelligence/adi-qwen3-8b-glm5.2-general-GGUF:Q4_K_M

Or download the .gguf and point any llama.cpp-based runtime at it.

What this model is

This is a knowledge distillation: a strong teacher (glm-5.2) generated high-quality answers across ~2,000 diverse general-knowledge prompts, and the Qwen3-8B student was fine-tuned to imitate them. The result reasons and responds noticeably more like its teacher on general topics, with more headroom than the 4B sibling, while staying small enough to run on a single consumer GPU.

What distillation does — and doesn't do. It transfers the teacher's reasoning style and answer quality, not net-new facts. An 8B model won't become an encyclopedia, though it carries more parametric knowledge than the 4B. For raw factual recall, retrieval-augmented generation (RAG) is the right tool, not fine-tuning. What you get here is an 8B that structures and explains like a much larger model on topics it already partly knows.

Training

Metric	Value
Training pairs	2,068
Teacher tokens generated	~1.36M
Epochs	3
Steps	777
LoRA rank / alpha	16 / 16
Precision	4-bit QLoRA
Hardware	single RTX 5060 Ti (16 GB)

The seed prompts were drawn from the human-written Databricks Dolly-15k dataset (filtered to remove items requiring an attached context passage, then deduplicated). The teacher was queried with thinking disabled so the student learns clean final answers rather than chain-of-thought.

Notes for re-builders

Qwen3-8B trains cleanly in 4-bit QLoRA via Unsloth — unlike the Qwen3.5 gated-delta/Mamba-hybrid layers, the standard Qwen3 architecture quantizes well for training. QLoRA on an 8B fits comfortably on a 16 GB card.
GGUF conversion was done with llama.cpp's convert_hf_to_gguf.py.

Intended use

General-purpose local assistant: explanations, reasoning, Q&A, and tool-calling workflows where a small, private, offline-capable model is preferred over a hosted API. Not intended as a source of authoritative facts without retrieval.

License

Apache-2.0, inherited from the Qwen3-8B base model. You are free to use, modify, and redistribute under the terms of that license. Distilled training data was generated using glm-5.2; users should review the teacher model's terms for their own use case.

Built at theLAB — Learning. Algorithms. Breakthroughs.