QW

RVN Qwen3.8 Flash Next Abliterated Uncensored

QwenGGUF by 0bserverxUpdated 2026-09-01
Download on Hugging Face
Parameters177B
Context256K
Layers48
Architectureqwen4exp
Smallest quant80.4 GB
Quick answer

RVN Qwen3.8 Flash Next Abliterated Uncensored needs about 120.5 GB of memory at Q4_K_M (a 119.2 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 18 quants, from 80.4 GB (Q2_K) to 167.6 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 3.2 GB.

190B Qwen4exp model with ablation edits and low refusal rate

RVN Qwen3.8 Flash Next Abliterated Uncensored is a 190 billion‑parameter Qwen4exp model. It is a heavily ablated derivative of the official Qwen3.8‑Flash‑Next checkpoint, with a measured 2 % strict refusal rate. The model is available in 18 quant families split into 144 shards.

Good at

  • Low refusal rate in instruction following
  • Large context support up to 262 144 tokens
  • High token throughput with Q6_K quant

Limits

  • Requires recent llama.cpp build with qwen4exp support
  • No vision projector included
  • Large memory footprint for high‑quality quants

Best for

  • Chat applications needing strict refusal control
  • Long‑form content generation with large context
  • Benchmarking Qwen4exp models on high‑end GPUs

Who it’s for. Developers and researchers who need a Qwen4exp model with a low refusal rate and large context length. The model is suitable for local deployment on systems with sufficient VRAM and RAM to handle the chosen quantization.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K80.4 GB (8 parts) 81.8 GB
Too big Get file →
IQ3_XS86.0 GB (8 parts) 87.3 GB
Too big Get file →
Q3_K_S88.8 GB (8 parts) 90.1 GB
Too big Get file →
IQ3_S88.9 GB (8 parts) 90.2 GB
Too big Get file →
IQ3_M89.5 GB (8 parts) 90.8 GB
Too big Get file →
Q3_K_M94.1 GB (8 parts) 95.4 GB
Too big Get file →
Q3_K_L96.7 GB (8 parts) 98.0 GB
Too big Get file →
IQ4_XS97.5 GB (8 parts) 98.8 GB
Too big Get file →
Q4_0100.0 GB (8 parts) 101.3 GB
Too big Get file →
IQ4_NL100.1 GB (8 parts) 101.4 GB
Too big Get file →
Q4_1111.0 GB (8 parts) 112.3 GB
Too big Get file →
Q4_K_S111.8 GB (8 parts) 113.1 GB
Too big Get file →
Q4_K_M119.2 GB (8 parts) 120.5 GB
Too big Get file →
Q5_0122.0 GB (8 parts) 123.3 GB
Too big Get file →
Q5_K_S127.7 GB (8 parts) 129.0 GB
Too big Get file →
Q5_1133.0 GB (8 parts) 134.3 GB
Too big Get file →
Q5_K_M134.1 GB (8 parts) 135.4 GB
Too big Get file →
Q6_K167.6 GB (8 parts) 168.9 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (≈81.8 GB memory at 8 k context) or IQ3_XS (≈87.3 GB). For 16 GB VRAM, Q3_K_S (≈90.1 GB) or IQ3_S (≈90.2 GB) are suitable. For 32 GB VRAM, Q4_0 (≈101.3 GB) or Q4_K_S (≈113.1 GB) provide higher quality while staying within budget. Memory estimates are approximate.

What GPU do you need for RVN Qwen3.8 Flash Next Abliterated Uncensored?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: no quant fits at an 8K context.

Which RVN Qwen3.8 Flash Next Abliterated Uncensored quant should you choose?

Q4_K_M is the usual starting point for RVN Qwen3.8 Flash Next Abliterated Uncensored: it balances file size and quality well. Here it is a 119.2 GB file.

If memory is tight, the smallest file is Q2_K at 80.4 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen4exp
Parameters177B (approximate, mixture of experts)
Layers48
Trained context262,144 tokens
KV cache (fp16)96 KB per token

Source

Quantized by0bserverx
FamilyQwen
Licenseother
Downloads (HF)22,911
Last updated2026-09-01

Frequently asked questions about RVN Qwen3.8 Flash Next Abliterated Uncensored

How much VRAM does RVN Qwen3.8 Flash Next Abliterated Uncensored need?

About 120.5 GB at Q4_K_M with an 8K context window. That is the 119.2 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run RVN Qwen3.8 Flash Next Abliterated Uncensored on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 81.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of RVN Qwen3.8 Flash Next Abliterated Uncensored GGUF?

It depends on the quant: Q2_K is 80.4 GB, IQ3_XS is 86.0 GB, Q3_K_S is 88.8 GB, IQ3_S is 88.9 GB, IQ3_M is 89.5 GB, Q3_K_M is 94.1 GB and 12 more.

What context length does RVN Qwen3.8 Flash Next Abliterated Uncensored support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.2 GB at 32K.

Where can I download RVN Qwen3.8 Flash Next Abliterated Uncensored GGUF?

The files are on Hugging Face in the 0bserverx/RVN-Qwen3.8-Flash-Next-Abliterated-Uncensored-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does RVN Qwen3.8 Flash Next Abliterated Uncensored use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run this model on 8 GB VRAM?

Use the Q2_K or IQ3_XS quant, which estimate around 81–87 GB memory at 8 k context. 8 GB VRAM is insufficient for any quant of this model.

Does this model support vision?

No. The repository excludes the vision projector and only contains the text/main model.

What llama.cpp build is required?

A recent build with qwen4exp support, such as commit eaf937655, is required to load these GGUF shards.

Other GGUFs of RVN-Qwen3.8-Flash-Next-Abliterated-Uncensored

Similar models

Browse all →

More from 0bserverx

All by 0bserverx →

More Qwen models

All Qwen →

Other large models

All large models →