Specification

Model name: Qwen3.8-27B-TURBO
Source repository: Psjuan/Qwen3.8-27B-TURBO
Base model: DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU
Publisher: Psjuan
Parameters: 27B
VRAM required: ~17.2 GB at Q4_K_M, 4k context (estimated) — runs on 24 GB cards
Hardware class: 25B – 50B — 48GB VRAM class
Licence: apache-2.0
Task: Image text to text
Context window: 256K tokens (as documented)
Ablation scope: Not stated by the publisher
Formats available: GGUF
First indexed: 8 Sept 2026
Last updated: 2 Nov 2026

Ablation technique

How Qwen3.8-27B-TURBO was modified, and what that implies.

Heretic

Automated directional ablation via the Heretic toolchain, which searches for the refusal direction and applies it with a KL-divergence budget so general capability is preserved.

Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.

From the publisher’s model card

“Important: This is the first fine tune to exceed 730 "arc-c" ("735": 144 pts higher than Qwen 3.8 27B) AND 880 ARC-E (The OpenAI, Claude and Gemini "zone of intelligence") in 8 bit and over 718 arc-c in 4 bit. This version is called TURBO because it drastically reduces thinking tokens (by 1/2 to as high as 1/10), yet...”

Running it

Commands are templates — confirm the exact repository and quant file before use.

llama.cpp / GGUF

hf download Psjuan/Qwen3.8-27B-TURBO --local-dir ./qwen3.8-27b-turbo
llama-cli -m ./qwen3.8-27b-turbo/<file>.gguf -p "..."

Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "Psjuan/Qwen3.8-27B-TURBO"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)

vLLM

vllm serve Psjuan/Qwen3.8-27B-TURBO --trust-remote-code

Downloads and variants (7)

Every published repository of Qwen3.8-27B-TURBO, including quantised re-releases by other authors.

Repository	Format	Downloads
Psjuan/Qwen3.8-27B-TURBOsource	GGUF	1.7K
bemematt/Qwen3.8-27B-TURBO	GGUF	1.6K
Jogmg/Qwen3.8-27B-TURBO-MAX	GGUF	1.5K
nimbowz/Qwen3.8-27B-TURBO-MTP-GGUF	GGUF	560
atesahmet0/Qwen3.8-27B-TURBO	GGUF	552
blastbeng/Qwen3.8-27B-TURBO	GGUF	342
Jfish123/Qwen3.8-27B-TURBO-16GB-GGUF	GGUF	134

Security-team assessment

Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.

Qwen3.8-27B-TURBO will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
Capability is inherited from DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU, not added by ablation — at 27B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
Declared licence is apache-2.0, but the terms that bind you are DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU’s — a re-release cannot grant more than its parent.

Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.