GILD verdict
Pick Qwen Code when your models are Qwen or other open models, on your hardware or Model Studio, or when a new team must start today.
Gemini CLI is the better choice for teams that hold a Code Assist subscription or Vertex AI billing and need its system settings file, until each contract ends and they move to Antigravity CLI. For 10 engineers, Coding Plan Pro costs about US$500 a month against US$190 for existing Code Assist Standard seats.
Qwen Code is the fork that went its own way, and Gemini CLI is the original Google is winding down. Qwen Code began from Gemini CLI v0.8.2, stopped syncing at v0.1, and now calls any model API or local model. Gemini CLI serves Code Assist subscribers who bought before October 9, 2026 and paid Gemini API and Vertex AI users, after Google moved free and Google AI plan users to Antigravity CLI on June 18, 2026 and stopped selling new Code Assist subscriptions on October 9, 2026.
Choosing between Qwen Code and Gemini CLI
Choose Qwen Code if
- Your models are Qwen, GLM, Kimi, or other open models, on your hardware or Alibaba Cloud Model Studio.
- You want an agent you can point at any OpenAI, Anthropic, Gemini, or Qwen API and switch with /model.
- A new team needs to start now, since Google sells no new Code Assist subscriptions.
Choose Gemini CLI if
- The team holds a Code Assist Standard or Enterprise subscription or bills Gemini through Vertex AI.
- An administrator must enforce the sign-in method, sandbox, and MCP allowlist from one system settings file.
- You plan the move to Antigravity CLI and want the original tool until each contract ends.
Side by side
| Dimension | Qwen Code | Gemini CLI |
|---|---|---|
| Plans and limits | Free agent; Coding Plan Pro US$50 a month with up to 6,000 requests per 5 hours, 45,000 a week, and 90,000 a month, interactive use only; pay-as-you-go keys or local models otherwise; the free Qwen OAuth tier ended April 15, 202636 | Existing Code Assist Standard subscribers get 1,500 requests per user per day and Enterprise 2,000; Google stopped selling new subscriptions on October 9, 2026; paid Gemini API keys and Vertex AI bill per token810 |
| Monthly cost for one workload (GILD calculation)GILD calculation from Alibaba Cloud's Coding Plan page and Google's Gemini for Google Cloud pricing on October 28, 2026. Inputs: 10 engineers, one Coding Plan Pro account each because accounts may not be shared, and Code Assist Standard at the one-year price. | For a team of 10 engineers, each on their own Coding Plan Pro account at US$50 a month: US$500 a month, with automated runs billed separately on pay-as-you-go keys6 | For a team of 10 engineers on existing Code Assist Standard subscriptions with a one-year commitment at US$19 per user per month: US$190 a month, available only to existing subscribers11 |
| Models | OpenAI, Anthropic, Gemini, and Qwen APIs, third-party routes such as DeepSeek, Kimi, and OpenRouter, or local models through Ollama or vLLM, switched with /model13 | Gemini models through a Code Assist sign-in, a Gemini API key, or Vertex AI; Google says Gemini CLI is not optimized for Gemini 4 Argon and points to Antigravity8 |
| Where it runs | Terminal on macOS, Linux, and Windows, plus a desktop app, web UI, SDKs, and chat channels1 | A Node.js terminal app for macOS, Windows, and Linux, preinstalled in Cloud Shell and Cloud Workstations, plus the run-gemini-cli GitHub Action12 |
| Integrations | VS Code, Zed, and JetBrains plugins, MCP servers, and an experimental daemon that serves other clients over ACP1 | MCP servers over stdio, SSE, or Streamable HTTP, with server allowlists set in configuration14 |
| Data use | The chosen route's terms govern prompts; usage statistics on by default for every auth method; Model Studio states it never trains on your data and stores call data47 | Paid and unpaid Gemini API services carry different terms, and Google uses content sent through unpaid services to improve its products9 |
| Admin controls | No documented SSO, SCIM, or audit log; OpenTelemetry events for each prompt and tool call can go to your own backend45 | A system settings file enforces the sign-in method, sandbox, and MCP allowlist; OpenTelemetry export for usage and tool calls1315 |
| License and lineage | Apache License 2.0; based on Gemini CLI v0.8.2 and developed independently since Qwen Code v0.112 | Apache License 2.0; Google is moving terminal users to Antigravity CLI, announced May 19, 20261617 |
| Choose it if | Choose Qwen Code if your models are Qwen or other open models, or you need to start a new team on a terminal agent today16 | Choose Gemini CLI if you hold a Code Assist subscription or Vertex AI billing and need its central settings until you move to Antigravity CLI1013 |
What practitioners report, side by side
Each product's figures come from its own practitioners, for the group you pick.
Overall signal
All practitioners rate Gemini CLI −3, 47 points above Qwen Code at −50. Most of the gap is in complaints, with 29 of Qwen Code's 42 sources negative against 22 of Gemini CLI's 64.
The score counts the authors from all practitioners that take a side, one vote each with their latest view, and is how far positive votes lead negative ones, per 100 votes, so Gemini CLI's −3 rests on votes split 20 positive, 20 mixed and 22 negative.
Each of Qwen Code's authors wrote one source here, so each source is one vote. A lead of a fifth of the votes reads positive and a lead of half strong positive, with the same lines below zero. The dots show each source, positive, mixed, negative. Use a task pilot to measure accepted work and review time.
Footprint on the Agentic Software Development stack
They differ on one of 12 layers. On CI/CD, Qwen Code touches it and Gemini CLI supports it. Coverage is each product's placement on this stack, as its product page shows it. Core is a job the product owns; Supports and Touches are jobs it helps with. Coverage says nothing about how well a product does them.
| Layer | Qwen Code | Gemini CLI |
|---|---|---|
| Ship | ||
| CI/CD | Touches | Supports |
The 11 layers where they match
| Layer | Qwen Code | Gemini CLI |
|---|---|---|
| Plan | ||
| Work planning | Supports | Supports |
| Build | ||
| Codebase context | Supports | Supports |
| Coding | Core | Core |
| Agent runtime | Core | Core |
| Tool execution | Core | Core |
| Validate | ||
| Debugging | Supports | Supports |
| Testing | Supports | Supports |
| Code review | Supports | Supports |
| Application security | Touches | Touches |
| Ship | ||
| Source control | Touches | Touches |
| Operate | ||
| Observability | Touches | Touches |
What practitioners praise
Among all practitioners, Gemini CLI alone has six praise themes that reach 5 sources from 3 authors. Open Source & Extensibility and Large Context lead, at 11 of 11 and 10 of 10. MCP & Editor Integrations and Harness for Local Models reach 5 sources from 3 authors only for Qwen Code, at 5 of 6 and 5 of 9. Gemini CLI needs a pilot on the same repository tasks before you choose on either.
| Praised for | Qwen Code | Gemini CLI |
|---|---|---|
| MCP & Editor Integrations | 5 of 6 | Not a theme in its sources |
| Harness for Local Models | 5 of 9 | Not a theme in its sources |
| Open Source & Extensibility | Not a theme in its sources | 11 of 11 |
| Large Context | Not a theme in its sources | 10 of 10 |
| Multi-Step Execution | Not a theme in its sources | 9 of 11 |
| Developer Experience | Not a theme in its sources | 7 of 20 |
| Speed | Not a theme in its sources | 4 of 7 |
| Headless Automation | Not a theme in its sources | 3 of 7 |
Each count is the sources that praise a theme, or raise it as a concern, out of those that discuss it. A row pairs the two products when both carry a theme under the same or an equivalent name.
A count ending in "so far" has under 5 sources from 3 authors among all practitioners. "Not a theme in its sources" means GILD found no theme of that name in that product's sources.
What practitioners raise as concerns
Among all practitioners, Gemini CLI alone has six concern themes that reach 5 sources from 3 authors. Reliability and Product Direction lead, at 25 of 27 and 14 of 16. Bugs & Stalled Tasks and Context & Cache Use reach 5 sources from 3 authors only for Qwen Code, at 16 of 16 and 5 of 6. Gemini CLI needs a pilot on the same repository tasks before you choose on either.
| Raised as a concern | Qwen Code | Gemini CLI |
|---|---|---|
| Bugs & Stalled Tasks | 16 of 16 | Not a theme in its sources |
| Context & Cache Use | 5 of 6 | Not a theme in its sources |
| Reliability | Not a theme in its sources | 25 of 27 |
| Product Direction | Not a theme in its sources | 14 of 16 |
| Quotas & Cost | Not a theme in its sources | 13 of 16 |
| Needs Close Steering | Not a theme in its sources | 7 of 7 |
| Requires Strong Review | Not a theme in its sources | 7 of 8 |
| Permissions & Safety | Not a theme in its sources | 7 of 8 |
17 sources
- QwenLM/qwen-code on GitHub (README)
- QwenLM/qwen-code LICENSE (Apache License 2.0)
- Qwen Code docs: Authentication
- Qwen Code docs: Terms of service and privacy notice
- Qwen Code docs: Telemetry
- Alibaba Cloud Model Studio: Coding Plan overview
- Alibaba Cloud Model Studio: Privacy notice
- Gemini CLI docs: Quota and pricing
- Gemini CLI docs: Terms and privacy
- Gemini Code Assist docs: Gemini Code Assist Standard and Enterprise subscriptions
- Gemini for Google Cloud pricing
- Gemini CLI docs: Installation
- Gemini CLI docs: Enterprise configuration
- Gemini CLI docs: MCP servers
- Gemini CLI docs: Observability with OpenTelemetry
- google-gemini/gemini-cli on GitHub
- Google Developers Blog: Transitioning Gemini CLI to Antigravity CLI (May 19, 2026)