LLM EXPLORER 58,608 MODELS INDEXED

Qwen3.5 9B Defiant Heretic NEO IMATRIX MAX MTP by ansulev

By ansulev · 24 downloads
HF Repo ↗ Copy URL

Qwen3.5 9B Defiant Heretic NEO IMATRIX MAX MTP is an open-source language model by ansulev. Features: 9b LLM, VRAM: 19.2GB, License: apache-2.0.

  Abliterated   All use cases   Bfloat16   Coder   Conversational   Creative   En   Endpoints compatible   Fiction   Fine tune   Heretic   Image-text-to-text   Mtp ggufs   Multi-stage tuned.   Neo imatrix   Qwen3.5   Qwen3 5   Reasoning   Region:us   Regular ggufs   Roleplaying   Safetensors   Sharded   Tensorflow   Thinking   Uncensored   Writing   Zh
Model Card on HF 🤗: https://huggingface.co/ansulev/Qwen3.5-9B-Defiant-Heretic-NEO-IMATRIX-MAX-MTP 

Qwen3.5 9B Defiant Heretic NEO IMATRIX MAX MTP Parameters and Internals

LLM NameQwen3.5 9B Defiant Heretic NEO IMATRIX MAX MTP
Repository 🤗https://huggingface.co/ansulev/Qwen3.5-9B-Defiant-Heretic-NEO-IMATRIX-MAX-MTP 
Model Name/mnt/d/Qwen3.5-9B-heretic-v2
Model Size9b
Required VRAM19.2 GB
Updated2026-08-17
Maintaineransulev
Model Typeqwen3_5
Model Files  4.9 GB: 1-of-4   5.0 GB: 2-of-4   5.0 GB: 3-of-4   4.3 GB: 4-of-4   0.1 GB
Supported Languagesen zh
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version5.12.1
Tokenizer ClassTokenizersBackend
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Qwen3.5 9B Defiant Heretic NEO IMATRIX MAX MTP

Best Alternatives
Context / RAM
Downloads
Likes
Ornith 1.0 9B0K / 18.8 GB2463295530
Qwable 9B Claude Fable 50K / 18.8 GB1088115
Qwen3.5 9B0K / 19.3 GB122077581799
Qwythos 9B Claude Mythos 5 1M0K / 18.8 GB37427906
Negentropy Claude Opus 4.7 9B0K / 18.8 GB6326
Qwen3.8 9B0K / 19.3 GB12841
Qwythos 9B V20K / 19.3 GB19208156
Qwen3.5 9B Base0K / 19.3 GB37734498
Tmax 9B0K / 17.9 GB2992615
Qwen3.5 9B NVFP40K / 11.2 GB2510812
Note: green Score (e.g. "73.2") means that the model is better than ansulev/Qwen3.5-9B-Defiant-Heretic-NEO-IMATRIX-MAX-MTP.