Even at stage 1, these versions are exceeding Qwen 3.8 core metrics on all fronts... and we are just getting started.
Now with three branches, Ultimate 1, 2 and 3.
ULT3
- has 3 components built and now in testing.
- more components being built.
- ULT1 and ULT2 will also share these (parts and assembled) too.
Metrics posted below. 2 huge "How to save the planet" examples (67 PAGES!) posted below too.
See also the SVG "Harley Pelican" below too.
BOOSTED:
- This model (and a few more) will soon have 12 reasoning and 12 instruct modes plus interactive help, model embedded system to select the best reasoning/instruct mode[s] for all use cases.
- Final testing is in progress (of this section) is in progress.
This is a "from scratch" branch and a split from this much larger project (10):
https://huggingface.co/DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU
THIS IS A WORK IN PROGRESS, estimated release mid/end of Oct.
If you sign up for access you will receive an email when final version(s) release.
This page will be updated regularly as the project/branch proceeds.
Prompt:
Generate an SVG of a pelican riding a Harley motorcycle, holding a shotgun with Rayban sun glasses on.
HTML [image capture, gen: Q4KS (4bit / 1/4 precision), non imatrix, xhigh from "Ultimate2" stage 1 branch ]:
NOTE: HTML to generate this is under the FILES section of the repo.
REG "PELICAN":
PROMPT: Generate an SVG of a pelican riding a bicycle.
HTML [image capture, gen: Q4KS (4bit / 1/4 precision), non imatrix, xhigh from "Ultimate2" stage 1 branch ]:
"ULTIMATE DETAILS":
- This branch starts with Qwen 3.8 27B and Qwen 3.6 Fable Fusion 711 (2300 likes, exceeds all benchs of all Qwen 3.6s with multiple 3rd party verifications).
- This branch is wholely and separate from Branch 1, 2 and "pre stage" models for these branchs.
- The goal of this branch is to very passively "curve" the thinking of Qwen 3.8 27B plus allow some of FF711 stronger characteristics to take hold.
- There are two working versions [early stages] with a reduction of 1/4 and 1/3 respectively VS "standard Qwen 3.8" with roughly 98% and 95% of thinking detail (including hyper detailed "math")/output detail and same or higher quality.
- These reductions are almost undetectable in thinking.
- Output quality has increased, and detail level has risen.
- Likewise output length (where required) as also lengthened.
- In other words the model does "xhigh" level reasoning in less tokens but delivers higher quality output with a stronger and more consise answer / longer answer rather than 16-32k of thinking with 4-8 k of output.
- Once we have some basic metrics [received now], we will begin precision (but passive) tuning -> ie improving decision making INSTEAD of directly tuning/re-tuning via reasoning datasets.
BRANCH 3 VS Branch 1/2:
- This branch is very different from Branch 1 and Branch 2 which use reasoning training (aggressive change).
- This branch is also a much longer, and detailed process and including multiple stages, tuning and "di-electric elbow grease".
- Heretic'ing / uncensoring will create a different branch at the end of this process and all other possible "vars" are also removed.
STAGE 1:
- Got some raw numbers / a place to start ; even at this early stage already exceeding Qwen 3.8 numbers by a large margin.
- Human testing evals in of changes between Ultimate 1 and 2 - "core" Qwen 3.8 reasoning intact, yet curve in thinking tokens/decision making is already showing.
- Influence of Qwen 3.6 Fable Fusion 711 is very clear.
- Assessing and prepping for raising core intelligence.
- May make 1 or more additional STAGE 1 branches.
- Branch 3 started [ultimate3] ; ground up build, piece by piece [this will be a slower start, gradual build].
STAGE 2:
- IN PROGRESS... drawing on FF771, Cold Fusion, TURBO and other experience to form a game plan.
- Ultimate 3, plus staged "pieces" building in progress.
arc/c arc/e boolq hswag obkqa piqa wino
[NEW BRANCH CORE START]
STAGE 1:
Ultimate 1 [1/4 reduction] - Primary base start.
mxfp8 0.645,0.844,0.902,...
Ultimate 2 [1/3 reduction] - Primary base start.
mxfp8 0.656,0.851,0.905,0.773,0.476,0.812,0.740
Ultimate 3 [Qwen Next Branch; ground up build.]
Qwen3.8-27B-ULT3-QNext-Stage1 => Qnext Reasoning Only Training Mild.
mxfp8 0.617,0.809,0.906,...
Qwen3.8-27B-ULT3-QNext-Stage1 => Qnext Reasoning Only Training Med.
- Training / benching in progress.
STAGE 2:
- in progress.
[QWENS]
[base, non heretic, untuned]
Qwen3.8-27B-Instruct:
mxfp8 0.591,0.782,0.896,0.746,0.448,0.801,0.711
mxfp4 0.581,0.771,0.889,0.738,0.442,0.798,0.713
Qwen3.6-27B-Instruct:
mxfp8 0.647,0.803,0.910,0.773,0.450,0.806,0.742
Qwen3.6-35B-A3B-Instruct
mxfp8 0.581,0.757,0.892,0.751,0.428,0.803,0.688
Qwen3.5-27B-Instruct:
mxfp8 0.557,0.711,0.868,0.533,0.452,0.706,0.695
NOTES:
- Models are tested in "Instruct" mode because this generally works better with the testing harness.
- Testing via "thinking" mode also shows the metrics (and changes) but not the true extent.
- In actual fact when the model IS in thinking mode, it will exceed INSTRUCT benchmark scores in most cases.
- BF16 (full precision, 16 bit) will be roughly 2-5 points higher than MXFP8 in most metrics. Some metrics may be slightly higher than this.
Qwen3.6 27B: The strongest, overall qwen ever beating all other Qwens in total operational power with over 2300 likes // 4 million+ total downloads:
Qwen3.8 27B: The highest scoring Qwen in brute, raw intelligence, using Qwen 3.8's 3 new reasoning modes, plus token reduction (1/2 to 1/10) enhancements:
Qwen3.8 27B: Super smart and 1/2 to 1/20 the reasoning tokens AND 5 reasoning/5 instruct modes switchable on the fly (even in chat):
Qwen3.8 27B: 99% power of BF 16 at 4 and 8 bit. Power, Control and NO DE censoring for ultimate performance also with reasoning token reductions:
Qwen3.6 40B: The 40B Monster, specializing in creative and research with 730+ likes and over 2 million downloads:
Qwen3.5 9B: At just 9B parameters it beats most untuned 27B models in both intelligence (640 ARC-C) and performance, plus features 5 reasoning and 5 instruct modes (Qwen 3.8) too:
STAGE 1 examples:
Q4ks, 4 bit (¼ precision), default Qwen parameters. NOTE: Some formatting lost on copy/paste.
Clear & very dry |████████████████████████████| ~80-120 W/m² Clear & moderate |██████████████████████| ~50-80 W/m² Partly cloudy |█████████████| ~20-40 W/m² Overcast |████| ~0-10 W/m²
Phase 2 (2030–2035): - Scale cool roofs and thermal mass - Expand urban districts - Radiative power pilots (WA, SA, NT) - Agricultural cooling (QLD, WA, SA, VIC)
Phase 3 (2035–2045): - National standards and codes - Large‑scale land‑surface projects (WA, NT, SA, QLD) - Integrated radiative cooling districts - Export of technology and best practice