Skip to content
Computer vision

Qwen3.8-27B-opencode MLX

a local coding agent for Mac

2 min read

This is an optimized version of the Qwen3.8-27B-opencode coding and terminal assistant, converted to run on Apple Silicon via the MLX framework. Trained with RL (AsyncGRPO) specifically for software engineering tasks and OpenCode agent workflows, the build packs both the text backbone and a built-in vision encoder alongside an MTP (multi-token prediction) layer for faster generation. A great pick for developers who want to run a capable local LLM without sacrificing performance.

Key features

  • Full compatibility with the MLX ecosystem for running on Apple Silicon Macs.
  • Strong coding and code analysis performance (SWE-bench Pro and Terminal-Bench benchmarks).
  • Support for processing images and screenshots via mlx-vlm.
  • Built-in MTP layer for faster token generation (speculative decoding).
  • Multiple quantization options (from 4-bit up to bf16) to match your available VRAM/unified memory.

What you can do

  • Spin up a local OpenAI-compatible server via mlx-lm to integrate with the OpenCode dev environment.
  • Write and test code directly inside Python scripts using a clean loading interface.
  • Debug terminal errors by feeding the model UI screenshots alongside a text prompt.
  • Speed up response generation by using the built-in MTP mechanism as a draft model.

Example

Launch a local server in the terminal to connect with OpenCode:

pip install -U "mlx-lm>=0.32.2"
mlx_lm.server --model netprojects/Qwen3.8-27B-opencode-MLX-bf16 --port 8000 \ --chat-template-args '{"enable_thinking": false}'
Read moreDetails
• SWE-bench Pro (hard 51): improved from 35.3% to 51.0% over the base model.
• Terminal-Bench 2.1: up from 46.4% to 52.4%.
• Model size: 55.6 GB in bf16; the recommended 6-bit version weighs in at 23.1 GB.
• License: Apache-2.0.
Discuss Telegram 0

Will it run?

Runs in 4-bit

Full precision does not fit, a quantised version does: quality barely changes down to 8 bits, more noticeably at 4 bits.

  • BF16weights 52 GB · ready files57 GB
  • FP8 / INT8recommendedweights 26 GB29 GB
  • 4-bitweights 15 GB17 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # netprojects/Qwen3.8-27B-opencode-MLX-bf16
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download netprojects/Qwen3.8-27B-opencode-MLX-bf16 --local-dir ~/models/Qwen3.8-27B-opencode-MLX-bf16
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

  1. Qwen3.8-27Bfamily root
  2. Qwen3.8-27B-opencode-adapted
  3. Qwen3.8-27B-opencode-MLX-bf16

New posts about versions of Qwen3.8-27B (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 2All on Hugging Face

About the model

netprojects/Qwen3.8-27B-opencode-MLX-bf16
Author
netprojects
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments