---

base_model:

• huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated

base_model_relation: quantized

license: other

license_name: qwen-community-1.0

library_name: gguf

pipeline_tag: image-text-to-text

tags:

• gguf
• qwen3.8-flash-next
• qwen4exp
• abliterated
• uncensored
• experimental
• agentionai

---

Qwen3.8-Flash-Next Gyro, abliterated by huihui-ai — GGUF (experimental)

The abliterated Qwen3.8-Flash-Next on a single GPU, in two sizes.

> Updated 2026-10-13: both files. Same file names, sizes and tensor types, better weights from an improved encode.

This is huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated

encoded the same way as our Gyro files. Abliteration

is huihui-ai's work: it removes the refusal direction from the model's weights. We only quantized it.

| File | For | GPU memory (64k context) | KLD vs the original model |

|---|---|---:|---:|

| Qwen3.8-Flash-Next-Gyro-S-abliterated-TQ1_0.gguf | 32 GB cards | 28.8 GiB | 0.397 |

| Qwen3.8-Flash-Next-Gyro-M-abliterated-TQ2_0.gguf | 48 GB: 2×24 GB cards, 64 GB unified-memory machines | ~37 GiB | 0.295 |

**Which setup fits your hardware, NVIDIA and AMD instructions, context limits, vision and speculative decoding are

all on the main Gyro card and apply unchanged here;

only the file names differ.**

Same architecture, size, speed and runtime as Gyro-S and Gyro-M. Abliteration changes 144 of the model's 1,224

weight tensors (the ones that write into the residual stream: attention and linear-attention outputs, and the down

projections of the shared and routed experts); everything else is byte-identical to the Gyro files. The TQ1_0 /

TQ2_0 suffix is a size class for the Hub's file browser, not the format: the experts use the trellis types TQK6/TQK7

(Gyro-S) and TQ2_T (Gyro-M), the same as the Gyro files. Upstream llama.cpp cannot load them yet; they run in

Strata 0.2.0 or newer (NVIDIA, AMD RX 7000 / RX 9000 / R9700),

halo-box/strix-llama.cpp master (AMD Strix Halo; CPU, Vulkan, HIP) and

agentionai/llama.cpp (CPU, Vulkan, CUDA).

Usage warning

Abliteration removes most of the model's refusals. Like huihui-ai's original, this model can produce harmful,

offensive or inappropriate content, and it is not suitable for public-facing or unsupervised use. You are

responsible for how you use it and for complying with the law where you are. See huihui-ai's model card for their

full warnings.

Quick start (Docker)

Our container has agentionai/llama.cpp ready (Vulkan; CUDA when built

yourself). Strata and strix-llama.cpp: see the Gyro card.

hf download agentionai/Qwen3.8-Flash-Next-Gyro-abliterated-Huihui-GGUF Qwen3.8-Flash-Next-Gyro-S-abliterated-TQ1_0.gguf \
  mtp-Qwen3.8-Flash-Next-draft.gguf mmproj-F16.gguf --local-dir ~/models

docker run --rm -it --device /dev/dri --group-add "$(getent group render | cut -d: -f3)" \
  -v ~/models:/models -p 8080:8080 agentionai/agention-llama:server \
  -m /models/Qwen3.8-Flash-Next-Gyro-S-abliterated-TQ1_0.gguf -ngl 999 -fa on --jinja --ngram-on-disk \
  -c 65536 -ctk q8_0 -ctv q8_0 --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0.05 \
  --mmproj /models/mmproj-F16.gguf

Or with -hf agentionai/Qwen3.8-Flash-Next-Gyro-abliterated-Huihui-GGUF:Gyro-S in a native build, which also fetches the

vision projector and, with --spec-type draft-mtp, the MTP draft. NVIDIA, vision on/off, speculative decoding and

memory per context length: exactly as on the Gyro card.

The MTP draft (updated 2026-10-05 to upstream llama.cpp's tensor names, weights unchanged; use agentionai/llama.cpp main

or the container from 2026-10-05 or later) was made from the original model; it still drafts well for the abliterated weights (81% of drafted

tokens accepted on a code prompt in a first check).

Quality

KL divergence and top-1 agreement, measured against the original (not abliterated) Qwen3.8-Flash-Next. These numbers therefore include the

abliteration itself, not only the quantization.

| File | GPU weights | KLD vs original | top-1 vs original | code KLD | code top-1 |

|---|---:|---:|---:|---:|---:|

| abliterated Gyro-S | 27.64 GiB | 0.397 | 75.3 % | 0.111 | 94.3 % |

| abliterated Gyro-M | 34.98 GiB | 0.295 | 78.6 % | 0.066 | 95.5 % |

Abliteration costs little here: the abliterated files are within 0.005 KL divergence of Gyro-S (0.401) and

Gyro-M (0.291). That is expected for technical text and code, which barely touch the refusal direction abliteration

removes. It says nothing about refusals themselves; for that, see huihui-ai's card.

Throughput

Same as the Gyro files, which have identical tensor types and sizes: Gyro-S on one AMD Radeon AI PRO R9700 32 GB,

prefill 1,246 / decode 57.8 tok/s; on AMD Strix Halo (Radeon 8060S), decode 31.7 tok/s; Gyro-M on Strix Halo, decode 26.4

(llama-bench, batch 1, no drafting). With strix-llama.cpp master (Vulkan) on Strix Halo: Gyro-S 700 / 36.1,

Gyro-M 738 / 30.1.

Status

Experimental, like the Gyro files: the format and kernels may change. Runs in agentionai/llama.cpp (Vulkan, CUDA), Strata and strix-llama.cpp; tested on AMD Strix Halo.

The weights come from our own encoder, which we are not publishing at this stage.

Changelog

• 2026-10-19: more engines. The files also run in Strata 0.2.0 and in

halo-box/strix-llama.cpp master. Files unchanged.

• 2026-10-13: both files updated. Same file names, sizes and tensor types; better weights from an improved encode.

Credits and license

Base model: Qwen/Qwen3.8-Flash-Next (Qwen community license 1.0, see LICENSE). Abliteration: huihui-ai. Quantization:

AgentionAI. Tensor format: trellis-coded quantization with Hadamard rotation, in the family of QTIP (Tseng et al., 2024). Encoder, trellis types and kernels: AgentionAI.

Support AgentionAI

These quants are released freely. If they save you VRAM or make Qwen more useful, you

can buy me a coffee or some GPU time and sponsor continued quantization and benchmarking on

GitHub. AgentionAI is a one-person team and can use your help.