BestLLMfor Your hardware. Your LLM. Your call.
The Local Copilot Kit APIOpen data Find my LLM
Model fiche

Qwen 3.8 27B Obliterated

By OBLITERATUS · China

Updated 2026-08-31

chat general multilingual
Parameters
28B
License
Apache 2.0
Context
256k
VRAM (Q4)
16 GB
Released
2026-08-19

Overview

An uncensored ('abliterated') fine-tune of Qwen 3.8 27B, dense architecture, with 262K context. Strips built-in refusal behavior for unrestricted local chat — runs at ~16GB VRAM in Q4.

When to pick this model

  • Local chat without built-in content restrictions
  • Multilingual conversation where safety filters get in the way
  • Long-context tasks needing 262K tokens
  • Commercial projects needing a permissive Apache 2.0 license

VRAM requirements by quantization

VRAM REQUIRED (GB)81216243248Q4_K_M16 GBQ5_K_M20 GBQ8_030 GBFP1656 GB
QuantizationVRAM required
Q4_K_M (recommended)16 GB
Q5_K_M20 GB
Q8_030 GB
FP16 (no quantization)56 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

In practice, Qwen 3.8 27B Obliterated needs a 16 GB card at Q4_K_M (16 GB). Stepping up to Q8_0 nearly doubles the footprint to 30 GB, and unquantized FP16 weights take 56 GB — plan your GPU around the Q4 or Q5 figure unless you specifically need the higher fidelity.

Without a GPU, Qwen 3.8 27B Obliterated needs roughly 36 GB of system RAM to run on CPU via llama.cpp or Ollama — workable for background jobs, but far slower than GPU inference. Throughput estimates from our compatibility engine: around 9 tokens/sec on entry-level GPUs, on the order of 14 tokens/sec on a mid-range card, and up to 22 tokens/sec on high-end hardware — assuming the chosen quantization fully fits in VRAM.

What hardware do you need

The table below matches Qwen 3.8 27B Obliterated to common GPU memory tiers, using the highest-fidelity quantization that fully fits each card class. Spilling layers to system RAM works but costs most of the speed, so size your card to the quantization you actually want to run.

GPU memoryExample cardsBest fit for Qwen 3.8 27B Obliterated
8 GBRTX 5070 Laptop, RTX 5060, RTX 5060 Ti 8GBDoes not fit — needs 16 GB at Q4_K_M
12 GBRTX 5070, RTX 5070 Ti Laptop, RTX 4080 LaptopDoes not fit — needs 16 GB at Q4_K_M
16 GBRTX 5080, RTX 4080 Super, Radeon RX 9070 XTQ4_K_M (16 GB used)
24 GBRTX 4090, Radeon RX 7900 XTX, RTX 5090 LaptopQ5_K_M (20 GB used)
32 GBRTX 5090Q8_0 (30 GB used)

Which hardware should you buy to run Qwen 3.8 27B Obliterated?

To run Qwen 3.8 27B Obliterated locally at Q4, you need ~16 GB of VRAM. The best value for this today is a RTX 5070 Ti 16GB (GIGABYTE Gaming OC) (16 GB VRAM).

Check RTX 5070 Ti 16GB (GIGABYTE Gaming OC) price on Amazon →

As an Amazon Associate, BestLLMfor earns from qualifying purchases, at no extra cost to you. It does not influence our independent rankings.

Strengths

  • Dense 28B general-purpose, multilingual
  • Native 262K context window
  • No built-in refusals — responds to prompts other models decline
  • Apache 2.0, fully permissive for commercial use

Limitations

  • Needs ~16GB VRAM at Q4 — 16-24GB GPU recommended
  • Abliteration can degrade alignment and safety guardrails
  • No official Ollama tag — install from Hugging Face

Typical workloads

In our catalog grid, Qwen 3.8 27B Obliterated is filed under Unfiltered Chat, Multilingual, Long Context 262K — the use cases where its size/quality trade-off makes the most sense. Its tags translate to concrete workloads: multilingual workloads.

The 256k-token context window is large enough to hold entire codebases' worth of files or long reports in a single prompt, which is what makes local RAG and document analysis practical. The Apache 2.0 license is permissive, so shipping it inside a commercial product raises no special legal questions.

Architecture & training

Architecture: Dense transformer · 28B parameters · 262K context · 'abliterated' variant (guardrails removed) of Qwen 3.8 27B

Training: Community fine-tune (OBLITERATUS) of Qwen 3.8 27B: abliteration intended to remove refusals. Official Qwen base; training details not published.

Verdict

A capable dense 28B multilingual chat model with the safety training stripped out — use deliberately, not by default.

Quick start

# HuggingFace : OBLITERATUS/Qwen3.8-27B-OBLITERATED

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.

Similar models worth comparing

Frequently asked questions

How much VRAM does Qwen 3.8 27B Obliterated need?

At the recommended Q4_K_M quantization, Qwen 3.8 27B Obliterated needs about 16 GB of VRAM. Q8_0 takes 30 GB, and unquantized FP16 weights take 56 GB.

Can Qwen 3.8 27B Obliterated run without a GPU?

Yes — with roughly 36 GB of system RAM it runs CPU-only through llama.cpp or Ollama. Expect a fraction of GPU speed, which is fine for background or batch jobs but slow for interactive chat.

What context window does Qwen 3.8 27B Obliterated support?

Qwen 3.8 27B Obliterated supports a 256k-token context window (262,144 tokens).

Can I use Qwen 3.8 27B Obliterated commercially?

Yes. Qwen 3.8 27B Obliterated is released under Apache 2.0, a permissive open-source license that allows commercial use, modification and redistribution.

How fast is Qwen 3.8 27B Obliterated on consumer hardware?

Our compatibility engine estimates on the order of 14 tokens/sec on a mid-range GPU and up to 22 tokens/sec on high-end cards, assuming the quantization fully fits in VRAM.

Which quantization of Qwen 3.8 27B Obliterated should I download first?

Start with Q4_K_M (16 GB) — the standard size/quality sweet spot. Step up to Q5_K_M or Q8_0 only if you have VRAM headroom. On a 24 GB card you can run up to Q5_K_M.

Tools

Is Qwen 3.8 27B Obliterated the right pick for you?

Compute self-hosted ROI → Back to catalog