Skip to main content
Qwen Code−50 Strong negative

GILD Review

GILD's call and the buying facts behind it.

From GILD research

Should we adopt Qwen Code?

Which model route will each engineer use for Qwen Code now that the free tier is gone?

  1. 30-Second TakePick Qwen Code if you run Qwen or other open models on your own hardware and want an open-source terminal agent.
  2. Best fitEngineers running Qwen or other open models locally through Ollama, vLLM, or LM Studio, and teams on Alibaba Cloud Model Studio, that want an open-source terminal agent with IDE plugins and MCP.GILD Review
  3. Consider another workflowIf you need SSO, SCIM, audit logs, or a vendor-backed seat plan, evaluate Claude Code or Codex. If your models are Gemini on Google Cloud billing, compare Gemini CLI, the project Qwen Code forked.GILD Review

Drawn from 53 sources and 3 operator insights

GILD analysis

Qwen Code Review

Practitioners rate Qwen Code highly as a harness for local Qwen models and report stalled tasks, auth failures, and heavy context use alongside. The Apache 2.0 agent calls Qwen, third-party, or local models, with IDE plugins and MCP, and Coding Plan Pro costs US$50 a month for interactive use. Adopt it with a chosen route per engineer, usage statistics off, and telemetry exported as your audit trail.

  • AI Coding Agents
  • Engineering Productivity

The decision

Best fit
Engineers running Qwen or other open models locally through Ollama, vLLM, or LM Studio, and teams on Alibaba Cloud Model Studio, that want an open-source terminal agent with IDE plugins and MCP.
Consider another workflow
If you need SSO, SCIM, audit logs, or a vendor-backed seat plan, evaluate Claude Code or Codex. If your models are Gemini on Google Cloud billing, compare Gemini CLI, the project Qwen Code forked.
Adoption condition
A sign-in route chosen per engineer, pay-as-you-go keys for anything automated, usage statistics turned off, and OpenTelemetry events exported to a backend you control.

Basis: the Qwen Code README and documentation and Alibaba Cloud's Coding Plan and privacy pages, fetched from October 28 to November 2, 2026, and 47 public accounts from people using Qwen Code. Qwen's own SWE-bench Verified figures are vendor-reported and not a GILD benchmark.

Buying & deployment factsWhat engineering leaders should verify before a broader rollout.
How you buy it

US$50 a month

Coding Plan Pro for interactive use; Qwen Code itself is free under the Apache License 2.0.

Source 1, Qwen Code team (Alibaba), Coding Plan: Coding PlanSource 2, Qwen Code team (Alibaba), Qwen Code README: Qwen Code README

Details

Qwen Code is free under the Apache License 2.0, and you pay for the model route. Alibaba Cloud Coding Plan Pro costs US$50 a month for interactive use in coding tools. Pay-as-you-go Model Studio keys, any third-party API, and local models through Ollama or vLLM also work. The Coding Plan Lite plan closed to new subscriptions on March 20, 2026.

Cost drivers

6,000 requests per 5 hours

Coding Plan Pro also caps 45,000 a week, and its key may not run automated scripts or backends.

Source 1, Qwen Code team (Alibaba), Coding Plan: Coding Plan

Details

Coding Plan Pro allows up to 6,000 requests per 5 hours, 45,000 a week, and 90,000 a month, and a task uses 5 to 30 or more model calls. Its key may not run automated scripts or application backends. Calls made with a general Model Studio key are billed pay-as-you-go on top of the plan.

Where it runs

Your machine, any model

A terminal agent with VS Code, Zed and JetBrains plugins; prompts go to the model route you set.

Source 2, Qwen Code team (Alibaba), Qwen Code README: Qwen Code READMESource 5, Qwen Code team (Alibaba), Authentication: AuthenticationSource 3, Qwen Code team (Alibaba), Terms and privacy notice: Terms and privacy notice

Details

A terminal agent for macOS, Linux, and Windows, with VS Code, Zed, and JetBrains plugins, MCP servers, a desktop app, a web UI, and SDKs. It calls OpenAI, Anthropic, Gemini, and Qwen APIs or local models through Ollama or vLLM, so prompts go to the route you configure. model.name sets the default model to activate when Qwen Code starts. With your own API key, your chosen provider's terms apply, and its documentation covers data usage and retention.

What to verify

Usage statistics on

On by default for every auth method, including your own key, until you turn them off.

Source 3, Qwen Code team (Alibaba), Terms and privacy notice: Terms and privacy noticeSource 4, Qwen Code team (Alibaba), Model Studio security certifications and privacy: Model Studio security certifications and privacySource 6, alibabacloud.com, ActionTrail supported services: ActionTrail supported servicesSource 7, qwenlm.github.io, qwen serve guide: qwen serve guideSource 8, alibabacloud.com, Alibaba Cloud: Model Studio product page: Alibaba Cloud: Model Studio product page

Details

Qwen Code sends anonymous usage statistics by default for every auth method, including your own key, until you turn them off. Alibaba says Model Studio has achieved SOC 2 compliance and does no unauthorized training on customer data. The qwen serve guide says to run one daemon per principal that needs its own audit log. The Qwen Code docs name no SSO, SAML or SCIM: GILD found 0 matches across 204 English docs and repository pages, checked Nov 2, 2026.

Pages GILD searched, by set, Nov 2, 2026
WherePages readMatches
Qwen Code documentation in English, every page in its sitemap2030
Qwen Code's GitHub repository page10
GILD search record, Nov 2, 2026

Sourced from Qwen Code and Alibaba Cloud documentation