--- license: mit base_model: - z-lab/Qwen3-Coder-Next-DFlash base_model_relation: quantized quantized_by: AlexAtomic pipeline_tag: text-generation library_name: gguf tags: - atomic-chat - dflash - speculative-decoding - draft-model - qwen - gguf - llama.cpp ---

Draft: z-lab/Qwen3-Coder-Next-DFlash · Target: Qwen/Qwen3-Coder-Next

**Qwen3-Coder Next DFlash**, the DFlash speculative-decoding **draft** converted to GGUF by [Atomic Chat](https://atomic.chat). Built straight from [z-lab](https://huggingface.co/z-lab)'s original weights. Runs fully offline. ## What this is [DFlash](https://github.com/z-lab/dflash) is a speculative-decoding method that drafts a whole **block** of candidate tokens in a single forward pass using a lightweight block-diffusion model, instead of one token at a time. This repo is the **draft component only** — it does nothing on its own. You run it alongside the target model **`Qwen/Qwen3-Coder-Next`**, which verifies the drafted block and keeps the longest correct prefix. Output is identical to running the target alone, just faster. > [!NOTE] > These GGUFs are **converted from z-lab's original weights**, not a repack of someone else's GGUF. The draft attaches to any GGUF of the target model (Atomic, unsloth, bartowski, ...). ## Run in llama.cpp Needs a build of [llama.cpp](https://github.com/ggml-org/llama.cpp) with DFlash speculative decoding (PR #22105). You supply the target as `-m` and this draft as `-md`: ```bash ./llama-server \ -m Qwen3-Coder-Next.gguf \ -md Qwen3-Coder-Next-DFlash.Q8_0.gguf \ --spec-type draft-dflash --spec-draft-n-max 15 \ -ngl 99 -fa on --jinja -c 8192 ``` DFlash is trained for **non-thinking** generation — pass `enable_thinking=false` in the chat template for best acceptance. ## Choosing a quant | Quant | Size | Notes | |---|---|---| | **`Q8_0`** | 0.51 GB | Recommended. Near-lossless draft head, small and fast to draft with. | ## Performance z-lab report up to **6.17x** lossless acceleration on their reference stack (vLLM / SGLang / Transformers). In `llama.cpp` today the DFlash port is newer: in our tests **dense** targets get roughly **1.8x-2.8x** end-to-end on code generation, and acceptance climbs on larger targets and structured/code output. Acceptance and speedup depend on the target and the content, not on the quantization. Speedups shrink on free-form prose and on small-active MoE targets. ## License Released by z-lab under the MIT license. Converted to GGUF by Atomic Chat. See the [DFlash paper](https://arxiv.org/abs/2602.06036) and [project page](https://github.com/z-lab/dflash).