supergemma4-26b-uncensored-v2
26BOtheruncensored
from google/gemma-4-26B-A4B-it
↓ 394K♥ 98442 repos
Qwythos-27B-v1
Qwythos-27B-v1 is a abliterated and uncensored variant of empero-ai/Qwythos-27B-v1, published on Hugging Face by empero-ai. It is 27B parameters, 48GB VRAM class class and apache-2.0 licence. It is distributed across 8 repositories in GGUF, Transformers, NVFP4 and GGUF (imatrix) formats, totalling 342.3K downloads.
Specification
- Model name
- Qwythos-27B-v1
- Source repository
- empero-ai/Qwythos-27B-v1
- Base model
- empero-ai/Qwythos-27B-v1
- Publisher
- empero-ai
- Parameters
- 27B
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Image text to text
- Context window
- 512K tokens (as documented)
- Ablation scope
- Not stated by the publisher
- Formats available
- GGUF, Transformers, NVFP4, GGUF (imatrix)
- First indexed
- 13 Jul 2026
- Last updated
- 21 Aug 2026
Ablation technique
How Qwythos-27B-v1 was modified, and what that implies.
Abliteration + uncensored fine-tune
Directional ablation combined with additional fine-tuning on unfiltered data, so behaviour diverges from the base model beyond refusal removal alone.
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
“Qwythos-27B-v1 is an open-weight, full-parameter reasoning model: the larger sibling of Qwythos-9B, trained on the exact same curriculum and sized up on a Qwen3.5-27B base. It is a complete pre-RL checkpoint, post-trained through SFT → DPO → ESFT.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download empero-ai/Qwythos-27B-v1-GGUF --local-dir ./qwythos-27b-v1
llama-cli -m ./qwythos-27b-v1/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "empero-ai/Qwythos-27B-v1"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve empero-ai/Qwythos-27B-v1 --trust-remote-codeDownloads and variants (7)
Every published repository of Qwythos-27B-v1, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| empero-ai/Qwythos-27B-v1-GGUF | GGUF | 334.3K |
| empero-ai/Qwythos-27B-v1source | Transformers | 4.5K |
| mradermacher/Qwythos-27B-v1-GGUF | GGUF | 1.9K |
| mradermacher/Qwythos-27B-v1-i1-GGUF | GGUF (imatrix) | 1.1K |
| Fastiraz/Qwythos-27B-v1-GGUF | GGUF | 260 |
| CocaKova/Qwythos-27B-v1-NVFP4 | NVFP4 | 140 |
| CrossNow/Qwythos-27B-v1 | Transformers | 11 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwythos-27B-v1 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 empero-ai/Qwythos-27B-v1, 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.
- This release adds fine-tuning on unfiltered data, so its behaviour diverges from the base model beyond refusal removal — differences you measure cannot be attributed to ablation alone.
- Declared licence is apache-2.0, but the terms that bind you are empero-ai/Qwythos-27B-v1’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 Qwythos-27B-v1 is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with empero-ai, and does not independently verify publisher claims. See responsible use.