What are the hardware requirements for Qwen3.5 35B A3B DFlash?

Qwen3.5 35B A3B DFlash requires approximately 0.8 GB of VRAM and supports a context window of 256K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Qwen3.5 35B A3B DFlash and how large is it?

Qwen3.5 35B A3B DFlash is developed by z-lab, a model with 35b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Qwen3.5 35B A3B DFlash perform on standard benchmarks?

Qwen3.5 35B A3B DFlash has the following published scores: LMSYS ELO 1396. Compare against reference models on this page or on the LLM Explorer leaderboards.

Qwen3.5 35B A3B DFlash by z-lab — VRAM 0.8GB, 256K context

Name: Qwen3.5 35B A3B DFlash
Author: z-lab

Qwen3.5 35B A3B DFlash is an open-source language model by z-lab. Features: 35b LLM, VRAM: 0.8GB, Context: 256K, License: apache-2.0, LLM Explorer Score: 0.41, ELO: 1396.

Arxiv:2602.06036 Base model:finetune:qwen/qwen3... Base model:qwen/qwen3.5-35b-a3... 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/Qwen3.5-35B-A3B-DFlash

Qwen3.5 35B A3B DFlash Benchmarks

LMSys ELO 1396 vs 1272 (so35) 9.7%

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

What is the LLM Explorer Rank (Score)

Qwen3.5 35B A3B DFlash Parameters and Internals

LLM Name	Qwen3.5 35B A3B DFlash
Repository 🤗	https://huggingface.co/z-lab/Qwen3.5-35B-A3B-DFlash
Base Model(s)	Qwen/Qwen3.5-35B-A3B Qwen/Qwen3.5-35B-A3B
Model Size	35b
Required VRAM	0.8 GB
Updated	2026-07-15
Maintainer	z-lab
Model Type	qwen3
Model Files	0.8 GB
Model Architecture	DFlashDraftModel
License	apache-2.0
Context Length	262144
Model Max Length	262144
Transformers Version	5.7.0
Vocabulary Size	248320

Rank the Qwen3.5 35B A3B DFlash capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation