← back to catalog · registered 2026-10-02 08:58

wy2pay/Qwen3.8-Flash-Next-abliterated-GGUF-IQ4XS

wy2pay GGUF second-order
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/wy2pay%2FQwen3.8-Flash-Next-abliterated-GGUF-IQ4XS"
Response includes
  • classification unknown
  • files 7
  • author_summary 1 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
?
Primary method

Unclassified

No clear signals of an abliteration technique in this model.
Confidence
UNKNOWN
Why this label 1 signal
No classification signals present. This may not be an abliterated model at all - it could be a repackaging, a merge with unrelated goals, or unrelated content that mentions the term.
  • no classification signals present (no abliterated, uncensored, or known producer/method markers)
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · 30-day
0
Likes
0
Model age
today
created 2026-10-02

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
apache-2.0
Tags
gguf llama.cpp qwen4_exp abliterated dgx-spark base_model:huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated base_model:quantized:huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated license:apache-2.0 endpoints_compatible region:us imatrix conversational
Total size
94.6 GB
Files
7
Quantizations
2
Registered
2026-10-02 08:58
Last updated on HF
2026-10-02 08:56

Files by quantization

F16 1 file 866 MB
mmproj-Qwen3.8-Flash-Next-Uncensored-F16.gguf 866 MB f0f352a9 download
Auxiliary files 6 files 94.6 GB
Qwen3.8-Flash-Next-Uncensored-IQ4_XS-00001-of-00003.gguf 41.7 GB 28a99b12 download
Qwen3.8-Flash-Next-Uncensored-IQ4_XS-00002-of-00003.gguf 41.7 GB 2a309e0b download
Qwen3.8-Flash-Next-Uncensored-IQ4_XS-00003-of-00003.gguf 7.42 GB ebc43c58 download
Qwen3.8-Flash-Next-Uncensored-MTP-draft.gguf 3.85 GB ff803fb4 download
README.md 2.36 KB 1319bf2a download
.gitattributes 1.92 KB 43967683 download

README current version from Hugging Face


license: apache-2.0
base_model: huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated
tags: [gguf, llama.cpp, qwen4_exp, abliterated, dgx-spark]

Qwen3.8-Flash-Next Abliterated — GGUF (IQ4_XS) for DGX Spark

GGUF ของ Qwen3.8-Flash-Next (125B / 6B-active MoE, arch qwen4_exp) ตัว abliterated (uncensored)
แปลงมาให้รันบน NVIDIA DGX Spark เครื่องเดียว (128GB unified memory)

ไฟล์

ไฟล์ ขนาด บทบาท
Qwen3.8-Flash-Next-Uncensored-IQ4_XS-0000{1,2,3}-of-00003.gguf 97.5 GB โมเดลหลัก (mirror จาก orcarouter/Qwen3.8-Flash-Next-Uncensored-GGUF)
Qwen3.8-Flash-Next-Uncensored-MTP-draft.gguf 4.14 GB MTP draft model สำหรับ speculative decoding
mmproj-Qwen3.8-Flash-Next-Uncensored-F16.gguf 0.91 GB vision projector (ถ้าใช้รูปภาพ)
lora-v1/ (เร็วๆ นี้) ~1 GB LoRA adapter จากการ SFT เพิ่ม (5 domains)
Qwen3.8-Flash-Next-lora-v1.gguf (เร็วๆ นี้) ~1-2 GB LoRA แบบ GGUF ใช้กับ --lora ได้ทันที

รันบน DGX Spark (llama.cpp)

hf download wy2pay/Qwen3.8-Flash-Next-abliterated-GGUF-IQ4XS

llama-server \
  -m Qwen3.8-Flash-Next-Uncensored-IQ4_XS-00001-of-00003.gguf \
  --model-draft Qwen3.8-Flash-Next-Uncensored-MTP-draft.gguf \
  --lora Qwen3.8-Flash-Next-lora-v1.gguf \
  -ngl 99 \
  --ctx-size 131072 \
  -ctk q8_0 -ctv q8_0 \
  --port 8080

หมายเหตุ:

  • IQ4_XS 97.5GB บนหน่วยความจำ 128GB (ใช้ได้จริง ~110GB) ต้องใช้ KV quant (q8_0)
    เพื่อเหลือที่ให้ context ยาว
  • context 1M tokens เต็ม ต้องลด quant เป็น IQ3_M หรือ KV q4
  • ต้องใช้ llama.cpp build ล่าสุดที่รองรับ qwen4exp

Provenance

Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.