Alibaba’s Qwen team has released Qwen Code v0.5.0 and paired it with a new VS Code Companion plugin, aiming to move the open-source coding agent beyond the terminal and into day-to-day IDE workflows.
What shipped in Qwen Code v0.5.0
Qwen Code v0.5.0 adds a tighter Visual Studio Code integration, including packaging work that bundles the CLI into the VS Code release package for smoother setup across operating systems.
The release also introduces a native TypeScript SDK intended to make Node.js/TypeScript integrations easier for developers building tools and automations around Qwen Code.
Another focus is continuity: Qwen Code v0.5.0 promotes “smart session management,” including automatically saving sessions so conversations can be resumed.
The update also expands model compatibility by adding support for “OpenAI-compatible reasoning models,” naming DeepSeek and Kimi among examples shared by the Qwen team.
In the v0.5.0 change log, Qwen Code also includes fixes and refactors tied directly to the VS Code IDE companion experience, including completion-menu adjustments and improvements in CLI detection and version management.
Key changes at a glance
| Area | What changed in v0.5.0 | Why it matters |
| VS Code integration | Bundles CLI into VS Code release package; improves prepackage cross-platform compatibility | Reduces friction in installing and running Qwen Code from VS Code across platforms |
| Developer SDK | Adds a native TypeScript SDK | Helps teams integrate Qwen Code into Node/TS tooling and pipelines |
| Sessions | Auto-save and resume workflow improvements | Enables longer-running tasks and “pick up where you left off” coding sessions |
| IDE companion fixes | Completion-menu and IDE companion improvements; CLI detection/version management refactors | Targets stability and usability inside the editor |
VS Code Companion plugin: what it does
Alongside the v0.5.0 release, QwenLM published “Qwen Code Companion” on the Visual Studio Marketplace as an extension designed to integrate Qwen Code directly with VS Code workspaces.
The Marketplace listing describes the extension as compatible with both VS Code and VS Code forks, signaling an effort to support a broader set of editor distributions.
Qwen’s own documentation describes the VS Code extension (Beta) as a “native graphical interface” that shows Qwen’s changes in real time within the IDE.
The docs list core features including a dedicated sidebar panel, conversation history access, and the ability to run multiple sessions simultaneously.
They also describe workflow controls such as “auto-accept edits mode,” designed to automatically apply Qwen’s edits as they are made.
Requirements and basic setup
Qwen’s VS Code integration guide lists a minimum requirement of VS Code 1.98.0 or higher.
It also instructs users to install the Qwen Code CLI with npm install -g qwen-code before using the extension.
The same guide directs users to install the extension from the Visual Studio Code Extension Marketplace.
| Step | Action | Notes |
| 1 | Install Qwen Code CLI (npm install -g qwen-code) | CLI is part of the documented VS Code setup flow |
| 2 | Install the VS Code extension from Marketplace | Qwen’s docs describe this as the standard install path |
| 3 | Confirm VS Code version | Docs require VS Code 1.98.0+ |
How this connects to Alibaba’s Qwen model stack
On GitHub, QwenLM describes Qwen Code as an open-source AI agent for the terminal, optimized for Qwen3-Coder, and positioned for tasks like understanding large codebases and automating development work.
The v0.5.0 direction—more IDE integration plus SDK support—fits that positioning by targeting how developers actually write, review, and refactor code inside editors rather than only in terminal sessions.
Qwen’s broader “coder” model family is also expanding: Qwen2.5-Coder is described as covering multiple sizes from 0.5B through 32B parameters, aimed at different deployment needs.
This matters because IDE assistants often need flexible model choices, ranging from smaller models that can run faster (or locally) to larger models that can reason over bigger tasks when infrastructure allows.
In parallel, Alibaba Cloud’s Model Studio documentation shows that VS Code-based coding extensions can connect to Qwen models through API-based workflows (for example, using the “Cline” extension with Qwen models from Model Studio).
That API-first approach is relevant because it reflects a growing pattern: model providers are competing not only on model quality, but also on how easily models plug into developer tools already used daily.
Why the v0.5.0 + VS Code release matters
The practical impact of this update is that Qwen Code is being pushed into a more “IDE-native” experience, where developers can review edits, manage files, and track conversation history without switching contexts.
For teams, multi-session support inside the editor can help separate tasks—such as refactoring, test generation, and debugging—into parallel threads without losing state.
And for tool builders, the native TypeScript SDK is a signal that Qwen Code is also targeting integration scenarios, not just end-user prompting.
From a usability perspective, the release notes’ emphasis on packaging and cross-platform compatibility suggests QwenLM is prioritizing setup reliability—often the difference between a tool being tried once versus adopted.
The explicit mention of “OpenAI-compatible reasoning models” also indicates Qwen Code is being shaped to operate in mixed-model environments where teams may switch models depending on cost, latency, or task complexity.
Release timeline and availability
QwenLM’s GitHub release tracking shows v0.5.0 as a published release in December 2025, with changes that include bundling the CLI into the VS Code release package and multiple VS Code companion improvements.
On the distribution side, the “Qwen Code Companion” listing appears on the Visual Studio Marketplace with a December 2025 publication date.
Qwen’s VS Code integration documentation positions the extension as “Beta” and outlines the baseline requirements and installation steps for developers adopting it now.
Final Thoughts
Qwen Code v0.5.0 and the VS Code Companion plugin show Alibaba’s Qwen team targeting real developer workflows—sessions, file handling, and in-editor visibility—rather than treating AI coding as a separate chat-only experience.
If the IDE companion continues to stabilize and the SDK matures, Qwen Code is set up to become not just a coding assistant, but a platform layer teams can embed into internal tooling.
The next practical signals to watch are how quickly the extension iterates beyond Beta expectations—especially around cross-platform setup, session reliability, and the depth of IDE actions it can safely automate.
