Qwen3.6-27B-Fable-5-Experimental-GGUF

Qwen 3.6 27B - Claude Fable 5 (Experimental)

<div style="border-left: 4px solid #D97706; background: rgba(217, 119, 6, 0.08); padding: 14px 16px; border-radius: 8px; margin: 16px 0;"> <strong style="color: #D97706;">Update · 6/22/2026</strong><br> GGUFs re-done with MTP built in.<br> <strong>llama.cpp MTP args:</strong> <code>--spec-type draft-mtp --spec-draft-n-max 3</code> </div>

Heres Qwen3.6 slightly over-trained on a very small dataset of fable 5 traces.

Had to use aggressive settings for the style transfer here due to the data contraints. Overall the model seems to be a better planner now and better at 3D modeling in three.js, making small games, and ML engineering.

Either way give it a shot, it really does look, talk, and plan like fable... but like all fine-tunes it comes with it's limitations.

Hopefully we can get some more well rounded data from the rest of the community to a do a much less aggressive tune on a larger dataset.

Reasoning was left untouched

Benchmarks

As always big thanks to @nightmedia for the speedy benchmarks.

                                 arc     arc/e	boolq
Qwen3.6-27B-Fable-5-Experimental 0.650	0.813	0.909
                     Qwen3.6-27B	0.637	0.798	0.911

Not really a benchmark but here's a procedurally generated duck that it zero-shotted lol (quant: q3_k_s)

---

The data for this model was easily extracted, formatted, and masked for training with Teich <img src="https://cdn-avatars.huggingface.co/v1/production/uploads/6837935ac3b7ffe0d2559ce9/-AxyvV4wfUY8uo87kNKkK.png" width="20" height="20" style="display: inline-block; vertical-align: middle; margin: 0 3px;">

This model was trained 2x faster with Unsloth and Huggingface's TRL library.

<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>