Qwen3 4B DFlash B16 is an open-source language model by z-lab. Features: 4b LLM, VRAM: 1.1GB, Context: 40K, License: mit, LLM Explorer Score: 0.3.
Arxiv:2602.06036 Custom code Dflash Diffusion Diffusion-language-model Efficiency Endpoints compatible Feature-extraction Flash-decoding Qwen Qwen3 Region:us Safetensors Speculative-decoding
Model Card on HF 🤗: https://huggingface.co/z-lab/Qwen3-4B-DFlash-b16
Qwen3 4B DFlash B16 Benchmarks
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 4B DFlash B16 Parameters and Internals
| LLM Name | Qwen3 4B DFlash B16 |
| Repository 🤗 | https://huggingface.co/z-lab/Qwen3-4B-DFlash-b16 |
| Model Size | 4b |
| Required VRAM | 1.1 GB |
| Updated | 2026-07-10 |
| Maintainer | z-lab |
| Model Type | qwen3 |
| Model Files | |
| Model Architecture | DFlashDraftModel |
| License | mit |
| Context Length | 40960 |
| Model Max Length | 40960 |
| Transformers Version | 4.57.3 |
| Vocabulary Size | 151936 |
Best Alternatives to Qwen3 4B DFlash B16
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Qwen3.5 4B DFlash | 256K / 1.3 GB | 18526 | 33 |
Note: green Score (e.g. "73.2") means that the model is better than z-lab/Qwen3-4B-DFlash-b16 .
Rank the Qwen3 4B DFlash B16 Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 54990 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a