language:
- en
- zh
license: apache-2.0
tags: - MTP
- fine tune
- heretic
- uncensored
- abliterated
- multi-stage tuned.
- all use cases
- thinking
- reasoning
- qwen3.6
- coder
- creative
- writing
- fiction
- roleplaying
- bfloat16
- all use cases
pipeline_tag: image-text-to-text
base_model: - DavidAU/Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP
library_name: transformers
For NEO Imatrix GGUFS - reg and MTP, as well as full model details/card please go here:
Model #2 of this release series is here (10 stages of "construction", this is a FURTHER modified version of "Qwen3.6-40B-Grand-Intelligence"):
https://huggingface.co/DavidAU/Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored
For NEO Imatrix GGUFS - reg and MTP, as well as full model details/card please go here:
"40B Eleanor" takes "40B-Grand-Intelligence" to the next level in terms of reduction of thinking tokens / quality and detail of output generation.
Qwen3.6-40B-Grand-Intelligence-Fable-Fusion-Uncensored-Heretic
40B : "There is a BIGGER storm coming... and it will take no prisoners."
1290 Tensors, 96 layers : 50% larger than 27B Qwen 3.6 it is based on.
We have selected 2 models for release, these 2 compliment each other in terms of benchs
and operations as well as thinking block/style and general overall performance.
- Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored ( Qwen3.6-40B-Grand-Intelligence-Six-711-717-BETA1-Tune W "The Deckard 40B" )
- Qwen3.6-40B-Grand-Intelligence-Fable-Fusion-Uncensored-Heretic ( Qwen3.6-40B-Grand-Intelligence-Six-711-717-BETA1-Tune )
These models contain MULTIPLE 711 and 717 Qwen 3.6 27B Fable Fusion models, with the "Deckard" version
containing all the unqiue "DNA" of the Deckard 40B model (previously released).
The "benches" far and above exceed Qwen 3.6 27B benches, and previously released 40B models by us.
Although the final benches are not as strong as 27B FF711, they are in the "ballpark" and human testing/results
show these models are stronger in some areas, which do not show up in the "general bench marking" results.
NOTE: #1 will have a separate repo :
https://huggingface.co/DavidAU/Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored
! Approaching release status ; there are 3 models being considered.
One is listed here (Qwen3.6-40B-Grand-Intelligence-Six-711-717-BETA1-Tune W "The Deckard 40B"), with 13k generation from testing (testing/benching still ongoing):
https://huggingface.co/DavidAU/Qwen3.6-40B-007-The-Deckard-FF
ABOUT:
40B version(s) based on the wildly powerful "711" (1500+ likes, 1.6 million+ downloads) which operates in/near "OpenAI, Claude and Gemini" closed source intelligence levels:
(Benchmarks of the "711" at the above repo that surpass the Qwen 3.6 27B in all respects)
Additional models used / being considered:
- The Deckard/Claude Opus 40B, 685+ likes, over 1 million downloads.
- 717 Unreleased "711" Fable fusion lab version.
- Other FF711 type / 40B Trained models from the lab.
TESTING:
- Qwen3.6-40B-Grand-Intelligence-Six-711-717-BETA1-Tune W "The Deckard 40B" ("Fable Fusion 711" Pipeline stage2 ) IN TESTING/BENCHING (two versions).
- Qwen3.6-40B-Grand-Intelligence-Six-711-717-BETA1-Tune ("Fable Fusion 711" Pipeline) IN TESTING/BENCHING.
- Finishing Alpha stage in 48 hrs then moving to BETA(s) (secondary stage post expand tuning).
- Qwen3.6-40B-Grand-Intelligence-Six-711-717-rawb (717 is an unreleased ARC-C 717 model.) now in Alpha Training.
- Qwen3.6-40B-Grand-Intelligence-Six-711-717-raw + rawb ; in testing.
- Grand Intelligence, Mix 5 Raw in testing.
- Prototype 1, light tune for testing of TWO 700+ ARC-C (OpenAi,Claude, and Gemini level intelligence) Qwen 3.6B super tunes joined "at the hip" to make a 40B monster.
- Prototype 4, Model "Fable-Fusion-711" joined "at the hip" with Deckard 27B (Qwen 3.6) to make a 40B monster in pre-tune testing.
- Testing / benching and post expansion tuning in progress.
IMPORTANT:
- THIS IS A WORK in PROGRESS, and will highlight some parts of this process (40B version of "711") as we proceed.
- NAME of this repo will CHANGE as the project proceeds.
- RUNNING Benchmarks [subject to change] below.
PROJECT NOTES:
- If you want to join the waitlist, you will be notified by email when the FINAL, fully tested and optimized version releases.
- During expansion is normal to lose some benchmarks levels, which are then restored during the post expansion tuning.
- It will (likely) take a number of tuning rounds/steps/stages to bring the 40B up to "700" club status. This can take several days due to training and eval time. We are only using local hardware.
RUNNING NOTES [reverse order]:
- BETA of Six-711-717-rawb and BETA Six-711-717-rawb [level 2] W The Deckard in testing/benching.
- Six-711-717-rawb -> Alpha Training, (complete), entering eval/benching and maybe multi-stage pipeline pending results.
- Six-711-717-raw / rawb in testing...
- Alpha 4 ("Qwen3.6-40B-Grand-Intelligence-Four-raw") in training/testing.
- Alpha 2 and 3, 3b in testing // additional "non trained" expanded also in testing.
- Selecting model(s) for Beta staging in progress.
- Prelim benchmarks for "Qwen3.6-40B-Grand-Intelligence-One-IQ-tune-alpha" (tuned expansion) posted, moving on to next stage(s). Other Alphas are pending too.
- Benchmark below for "Qwen3.6-40B-Grand-Intelligence-One" (the root model for "Qwen3.6-40B-Grand-Intelligence-One-IQ-tune-alpha") BEFORE post expansion (27B to 40B) training.
BENCHMARKS by Nightmedia
arc/c arc/e boolq hswag obkqa piqa wino
[#1] "Qwen3.6-40B-Fable-Fusion-6-Core-Deckard-Eleanor-Heretic-Uncensored"
Qwen3.6-40B-Grand-Intelligence-Six-711-717-BETA1-Tune
W "The Deckard Claude-Opus 40B" ("Fable Fusion 711" Pipeline stage 2 )
mxfp8 0.687,0.857,0.908,0.825,0.500,0.818,0.771
REMARKS:
Off the scale detail. Stable. Intelligent, never seen before
generation quality/detail. Depth of thought too.
VASTLY reduced thinking tokens/thinking block ; auto-variable.
(1/10 to 1/2 the number of thinking tokens VS "norm" Qwen)
[#2] "Qwen3.6-40B-Grand-Intelligence-Fable-Fusion-Uncensored-Heretic"
(Qwen3.6-40B-Grand-Intelligence-Six-711-717-BETA1-Tune)
[FF711 Pipeline]
mxfp8 0.698,0.862,0.904
REMARKS: Strong generation, Strong Intelligence, High detail. Stable.
Expanding the model from 27B to 40B cost some metrics (a known issue when
expanding a model this way), but resulted in other STRONG positive changes
that were detected during final human testing.
---
Qwen3.6-40B-Grand-Intelligence-Six-711-717-rawb-Alpha-Tune
[complex tune, mid, Heretic 2.0]
mxfp8 0.692,0.859,0.903,...
REMARKS: Strong, close to perfect at this stage.
Qwen3.6-40B-Grand-Intelligence-One-IQ-tune-alpha4
[complex tune, mid]
mxfp8 0.675,0.856,0.906,...
REMARKS: 711/Deckard mix. Stable. But metrics drop too far at this stage.
Qwen3.6-40B-Grand-Intelligence-One-IQ-tune-alpha3b
[mid-light repair, deeper tune only]
mxfp8 0.695,0.864,0.902,0.819,0.494,0.814,0.772
REMARKS: Slightly lower, however 100% stable. SOTA IQ/power at 40B.
Qwen3.6-40B-Grand-Intelligence-One-IQ-tune-alpha3
[mid-light repair tune only]
mxfp8 0.701,0.862,0.903,...
REMARKS: Excellent, but unstable with some prompts.
Qwen3.6-40B-Grand-Intelligence-One-IQ-tune-alpha2
[light repair tune only]
mxfp8 0.690,0.864,0.908,...
Qwen3.6-40B-Grand-Intelligence-One-IQ-tune-alpha
[light repair tune only]
mxfp8 0.689,0.859,0.903,...
------------------------------------------------------------
RAW EXPANDED MODEL(s) - pre stage to be tuned/adjusted.
------------------------------------------------------------
Qwen3.6-40B-Grand-Intelligence-Six-711-717-rawb
mxfp8 0.677,0.858,0.893
Qwen3.6-40B-Grand-Intelligence-Six-711-717-raw
[NOT TRAINED YET, expansion only]
mxfp8 0.673,0.856,0.894
Qwen3.6-40B-Grand-Intelligence-Four-raw
[NOT TRAINED YET, expansion only]
mxfp8 0.679,0.853,0.905
Qwen3.6-40B-Grand-Intelligence-One
[NOT TRAINED YET, expansion only]
mxfp8 0.675,0.860,0.900,0.795,0.478,0.801,0.752
------------------------------------------------------------
ORG MODELS FROM QWEN, no tuning, non heretic.
------------------------------------------------------------
Qwen3.6-27B-Instruct: [base, non heretic]
mxfp8 0.647,0.803,0.910,0.773,0.450,0.806,0.742
Qwen3.6-35B-A3B-Instruct [base, non heretic]
mxfp8 0.581,0.757,0.892,0.751,0.428,0.803,0.688
Qwen3.5-27B-Instruct: [base, non heretic]
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.