LLM EXPLORER 60,702 MODELS INDEXED

Qwen3.6 35B A3B RFT by shaneMattner

By shaneMattner · 13 downloads
HF Repo ↗ Copy URL

Qwen3.6 35B A3B RFT is an open-source language model by shaneMattner. Features: 35b LLM, VRAM: 69.3GB, Context: 256K, License: apache-2.0, LLM Explorer Score: 0.23.

  Arxiv:2106.09685   Arxiv:2604.01193 Base model:finetune:qwen/qwen3... Base model:qwen/qwen3.6-35b-a3...   Bf16   Code-generation   Coding   Conversational   Deltanet   En   Endpoints compatible   Linear-attention   Lora   Lora-merged   Model-index   Moe   Qwen   Qwen3.6   Qwen3 5 moe   Region:us   Rejection-fine-tuning   Safetensors   Self-distillation   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/shaneMattner/Qwen3.6-35B-A3B-RFT 

Qwen3.6 35B A3B RFT Parameters and Internals

LLM NameQwen3.6 35B A3B RFT
Repository 🤗https://huggingface.co/shaneMattner/Qwen3.6-35B-A3B-RFT 
Base Model(s)  Qwen/Qwen3.6-35B-A3B   Qwen/Qwen3.6-35B-A3B
Model Size35b
Required VRAM69.3 GB
Updated2026-08-07
MaintainershaneMattner
Model Typeqwen3_5_moe
Model Files  49.7 GB: 1-of-2   19.6 GB: 2-of-2
Supported Languagesen
Model ArchitectureQwen3_5MoeForCausalLM
Licenseapache-2.0
Context Length262144
Model Max Length262144
Transformers Version5.8.0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size248320
LoRA ModelYes
Errorsreplace

Best Alternatives to Qwen3.6 35B A3B RFT

Best Alternatives
Context / RAM
Downloads
Likes
Qwen 35B A3B SignOfFour Coder256K / 70.6 GB17223
UltiMerge256K / 69.3 GB5572
SignOfFour256K / 70.6 GB7281
DAOS Fusion256K / 69.3 GB342
Qwen AgentWorld 35B A3B OQ40K / 20.3 GB22937
Qwen3.6 28B0K / 56.5 GB9534
Ornstein3.6 35B A3B RYS SABER256K / 71 GB1799
Ornstein3.6 35B A3B RYS SABER256K / 71.1 GB1129
Note: green Score (e.g. "73.2") means that the model is better than shaneMattner/Qwen3.6-35B-A3B-RFT.