LLM EXPLORER 59,766 MODELS INDEXED

Qwen3.5 4B DFlash by z-lab

By z-lab · 18526 downloads
HF Repo ↗ Copy URL

Qwen3.5 4B DFlash is an open-source language model by z-lab. Features: 4b LLM, VRAM: 1.3GB, Context: 256K, License: apache-2.0, LLM Explorer Score: 0.43, ELO: 1397.

  Arxiv:2602.06036 Base model:finetune:qwen/qwen3...   Base model:qwen/qwen3.5-4b   Block-diffusion   Custom code   Dflash   Diffusion-language-model   Draft-model   Efficiency   Feature-extraction   Qwen   Qwen3   Qwen3.5   Region:us   Safetensors   Sglang   Speculative-decoding   Speculative-decoding-draft
Model Card on HF 🤗: https://huggingface.co/z-lab/Qwen3.5-4B-DFlash 

Qwen3.5 4B DFlash Benchmarks

LMSys ELO 1397 vs 1272 (so35) 9.8%
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

Qwen3.5 4B DFlash Parameters and Internals

LLM NameQwen3.5 4B DFlash
Repository 🤗https://huggingface.co/z-lab/Qwen3.5-4B-DFlash 
Base Model(s)  Qwen/Qwen3.5-4B   Qwen/Qwen3.5-4B
Model Size4b
Required VRAM1.3 GB
Updated2026-07-11
Maintainerz-lab
Model Typeqwen3
Model Files  1.3 GB
Model ArchitectureDFlashDraftModel
Licenseapache-2.0
Context Length262144
Model Max Length262144
Transformers Version5.7.0
Vocabulary Size248320

Best Alternatives to Qwen3.5 4B DFlash

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3 4B DFlash B1640K / 1.1 GB2329629