← back to catalog · registered 2026-09-18 12:56

com-kotobalabs/Nex-N2.5-mini-Uncensored-W4A16-AutoRound

com-kotobalabs MoE second-order
curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/com-kotobalabs%2FNex-N2.5-mini-Uncensored-W4A16-AutoRound"
Response includes
  • classification m1
  • files 16
  • author_summary 3 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · 30-day
0
Likes
0
Model age
today
created 2026-09-18

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
apache-2.0
Tags
safetensors qwen3_5_moe auto-round w4a16 int4 gptq vllm uncensored abliterated text-generation conversational base_model:orcarouter/Nex-N2.5-mini-Uncensored

Related

Total size
19.2 GB
Files
16
Quantizations
1
Registered
2026-09-18 12:56
Last updated on HF
2026-09-18 12:08

Files by quantization

Auxiliary files 16 files 19.3 GB
model-00004-of-00005.safetensors 4.00 GB 083566a6 download
model-00003-of-00005.safetensors 4.00 GB 14b31070 download
model-00001-of-00005.safetensors 4.00 GB ecc5fa4a download
model-00002-of-00005.safetensors 4.00 GB 9dd6d8f5 download
model-00005-of-00005.safetensors 3.22 GB d32e14a8 download
tokenizer.json 19.1 MB 87a7830d download
model.safetensors.index.json 12.6 MB e77a8660 download
config.json 33.9 KB e27c14e3 download
quantization_config.json 28.9 KB 81c31591 download
chat_template.jinja 7.56 KB 34398e43 download
README.md 2.54 KB 122eb287 download
.gitattributes 1.60 KB a09db2ea download
processor_config.json 1.29 KB a9804dd9 download
tokenizer_config.json 1.14 KB 5344df8c download
preprocessor_config.json 478 B d7c5bf00 download
generation_config.json 137 B 637943d6 download

README current version from Hugging Face


license: apache-2.0
base_model: orcarouter/Nex-N2.5-mini-Uncensored
tags:

  • auto-round
  • w4a16
  • int4
  • gptq
  • vllm
  • qwen3_5_moe
  • uncensored
  • abliterated
    pipeline_tag: text-generation

Nex-N2.5-mini-Uncensored-W4A16-AutoRound

W4A16 (INT4 weights, group size 128, symmetric, BF16 activations) quantization of
orcarouter/Nex-N2.5-mini-Uncensored
(the abliterated Nex-N2.5-mini, a Qwen3.5-MoE 35B-A3B vision-language model) in
auto_gptq format, loadable by vLLM (Marlin / XPU) and SGLang.

Why this exists

orcarouter publishes the Uncensored weights as BF16 (65 GiB), FP8 (34 GiB, Hopper+),
NVFP4 (22 GiB, Blackwell only), MLX (Apple) and GGUF (llama.cpp). None of those loads
in vLLM on a 32 GiB non-Blackwell card. [quant-mind/Nex-N2.5-mini-W4A16-AutoRound]
(https://huggingface.co/quant-mind/Nex-N2.5-mini-W4A16-AutoRound) filled that gap for
the STOCK model; this repository applies the same published recipe to the Uncensored
weights, so the two differ in base weights only. Built for the murakumo fleet's
Intel Arc Pro B70 (32 GiB) head, where vLLM XPU + XPU graph measured ~1,128 tok/s
aggregate on the stock W4A16 checkpoint vs ~137 tok/s for llama.cpp IQ4_XS on the same
card (root ADR-2609181615, com-junkawasaki/root).

Recipe (identical to quant-mind's config.json)

  • Intel AutoRound 0.15.1, --scheme W4A16 --group_size 128 (sym), --iters 200,
    --nsamples 128 --seqlen 2048, calibration NeelNanda/pile-10k, --low_gpu_mem_usage
  • Kept in BF16: mlp.gate (router), mlp.shared_expert_gate, mtp, lm_head,
    embeddings, the visual tower
  • Quantized: linear_attn.*_proj, self_attn.{q,k,v,o}_proj, all 256 mlp.experts.*,
    mlp.shared_expert.*
  • Export format: auto_gptq
  • Built on Modal (H100) by the job tools/modal-quant/nex_n25_mini_uncensored_w4a16.py
    in kotoba-lang/murakumo.

Use

vllm serve com-kotobalabs/Nex-N2.5-mini-Uncensored-W4A16-AutoRound     --reasoning-parser qwen3 --max-model-len 65536 --gpu-memory-utilization 0.90

The model thinks before answering; without --reasoning-parser qwen3 the reasoning
leaks into content.

Caveats

  • Abliterated / uncensored derivative: it will not refuse. Deploy behind your own policy
    layer; the murakumo gateway runs it only behind governed organisms.
  • Quality vs the BF16 source is not evaluated here beyond a coherence probe; the
    quantization error profile is the one AutoRound publishes for this recipe.
  • License follows the base: Apache-2.0 (nex-agi/Nex-N2.5-mini) as relicensed by orcarouter.
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in app" button that hands off directly to a local runtime of your choice - Infrahuman, LM Studio, or Ollama. No API keys, no subscription, no prompt leakage.