LLM EXPLORER 67,711 MODELS INDEXED

Qwen3.5 4B Qwen3.8 27B Draft by nvidia

By nvidia · 290 downloads

Qwen3.5 4B Qwen3.8 27B Draft is an open-source language model by nvidia. Features: 27b LLM, VRAM: 8.4GB, Context: 256K, License: other.

Base model:finetune:qwen/qwen3...   Base model:qwen/qwen3.5-4b   Conversational   Draft   Model optimizer   Modelopt   Nvidia   Qwen3.5   Qwen3.8   Qwen3 5 text   Region:us   Safetensors

Qwen3.5 4B Qwen3.8 27B Draft Parameters and Internals

LLM NameQwen3.5 4B Qwen3.8 27B Draft
Repository πŸ€—https://huggingface.co/nvidia/Qwen3.5-4B-Qwen3.8-27B-Draft 
Base Model(s)  Qwen/Qwen3.5-4B   Qwen/Qwen3.8-27B   Qwen/Qwen3.5-4B   Qwen/Qwen3.8-27B
Model Size27b
Required VRAM8.4 GB
Updated2026-11-18
Maintainernvidia
Model Typeqwen3_5_text
Model Files  8.4 GB
Model ArchitectureQwen3_5ForCausalLM
Licenseother
Context Length262144
Model Max Length262144
Transformers Version5.14.1
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size248320
Errorsreplace

Best Alternatives to Qwen3.5 4B Qwen3.8 27B Draft

Best Alternatives
Context / RAM
Downloads
Likes
Hemmingway 1256K / 53.8 GB12371880
Hemmingway 1256K / 53.8 GB9089888
JEV 27B256K / 53.9 GB17864379
OrcaSAQ 2 27B256K / 12.3 GB1330169
Diffusion Qwen3.8 27B256K / 53.8 GB186029
Qwen3.6 27B OBLITERATED256K / 53.6 GB15137184
Rewrite2 27B Bf16256K / 53.7 GB4213
Darwin 27B RSI256K / 53.8 GB45914
...2 Abliterated OptiQ 3.7bpw Mlx256K / 13.4 GB672714
Gork 27B Cybersecurity256K / 54.7 GB571
Note: green Score (e.g. "73.2") means that the model is better than nvidia/Qwen3.5-4B-Qwen3.8-27B-Draft.