---
title: "Edit completion works with Qwen3.6 35B A3B!"
url: https://daily.dev/posts/edit-completion-works-with-qwen3-6-35b-a3b--db8tipejo
source_url: http://www.randomhacks.net/2026/05/13/edit-completion-with-qwen36
type: article
source: "Planet Haskell"
published: 2026-05-13T07:00:17.168Z
updated: 2026-05-14T02:13:57.321Z
tags: ["llama-cpp", "local-ai", "qwen", "zed"]
reading_time: 2
upvotes: 0
comments: 0
language: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Edit completion works with Qwen3.6 35B A3B!

**[Planet Haskell](https://daily.dev/sources/phaskell)** · 2 min read · 0 upvotes · 0 comments

## Summary

A short guide on setting up Qwen3.6 35B A3B as a local FIM (fill-in-the-middle) autocomplete model using llama-server and the Zed editor. The post covers the llama-server launch command with recommended flags for VRAM efficiency, and the Zed configuration needed to enable edit predictions via the OpenAI-compatible API. The author notes completions work but tend to be too long due to limited post-processing in Zed, suggesting a proxy layer could improve results.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <http://www.randomhacks.net/2026/05/13/edit-completion-with-qwen36>

## Similar posts on daily.dev

- [Lab notebook: Edit completion \#1](https://daily.dev/posts/lab-notebook-edit-completion-1-izb8vwytl) · Planet Haskell · 0 upvotes · 0 comments

---

Tags: [#llama-cpp](https://daily.dev/tags/llama-cpp), [#local-ai](https://daily.dev/tags/local-ai), [#qwen](https://daily.dev/tags/qwen), [#zed](https://daily.dev/tags/zed)

[View this post on daily.dev](https://daily.dev/posts/edit-completion-works-with-qwen3-6-35b-a3b--db8tipejo)