⌦abliteratedmodels.org

Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano

Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano is a refusal-ablated variant of prithivMLmods/Qwen3.6-35B-A3B-abliterated-MAX, published on Hugging Face by Hyphonical. It is 35B total parameters (mixture-of-experts), 48GB VRAM class class and apache-2.0 licence. It is distributed across 2 repositories in GGUF format, totalling 5.8K downloads.

Abliteration (directional ablation)35B MoEQwen25B – 50B↓ 5,766♥ 3
View on Hugging Face ↗Base model ↗

Specification

Model name
Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano
Source repository
Hyphonical/Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano-GGUF
Base model
prithivMLmods/Qwen3.6-35B-A3B-abliterated-MAX
Publisher
Hyphonical
Parameters
35B (mixture-of-experts)
VRAM required
~22.1 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
Text generation
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
GGUF
First indexed
1 May 2026
Last updated
1 May 2026

Ablation technique

How Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano was modified, and what that implies.

Abliteration (directional ablation)

The single residual-stream direction that mediates refusal is identified from harmful/harmless prompt pairs and projected out of the model weights.

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

“This model was quantized using apex-quant with the i-nano profile and an importance matrix calibrated on a diverse code/math/reasoning dataset.”

Running it

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

llama.cpp / GGUF
hf download Hyphonical/Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano-GGUF --local-dir ./qwen3.6-35b-a3b-abliterated-max-apex-i-nano
llama-cli -m ./qwen3.6-35b-a3b-abliterated-max-apex-i-nano/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "Hyphonical/Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve Hyphonical/Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano-GGUF --trust-remote-code

Downloads and variants (1)

The only published repository of Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano.

RepositoryFormatDownloads
Hyphonical/Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano-GGUFsourceGGUF5.8K

Security-team assessment

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

  • Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano 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 prithivMLmods/Qwen3.6-35B-A3B-abliterated-MAX, not added by ablation — at 35B 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 prithivMLmods/Qwen3.6-35B-A3B-abliterated-MAX’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.

Metadata for Qwen3.6-35B-A3B-abliterated-MAX-APEX-i-nano is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with Hyphonical, and does not independently verify publisher claims. See responsible use.