Gemini CLI alternatives in 2026: 8 coding agents to use instead

Disclosure: Enjoy is our product, and it runs Antigravity and OpenCode. We checked every option below against its maker’s own pages on November 5, 2026, and we say where another one is the better pick, Gemini CLI included.

Gemini CLI alternatives compared

Agent	Best for	Price, checked November 5, 2026	Runs on	Gemini?
Antigravity CLI	The closest move, on your Google account	Free; Google AI Pro $19.99 a month; Ultra $99.99 or $199.99	Mac, Windows, Linux	Gemini 3.8 Flash free; Gemini 4 Argon from paid Pro
OpenCode	Any model, from a free start	Free; Go $10 a month; Go Plus $40; other providers at their prices	Mac, Windows, Linux	With your API key
Cline	Your own key, in VS Code or a terminal	Free; ClinePass $9.99 a month; inference at cost	Terminal, VS Code; desktop app for Mac and Windows	With your API key
Qwen Code	A fork of Gemini CLI, for many providers	Free; its plans and API keys at their prices	Mac, Windows, Linux	With your API key
Aider	Pair programming in Git, on your own key	Free; models at their providers’ prices	Mac, Windows, Linux	With your API key
Codex CLI	OpenAI’s models in a terminal	ChatGPT Plus $20 a month; Pro from $100; or an API key	Mac, Windows, Linux	Not on a ChatGPT plan
Claude Code	Claude, from the company that makes it	Claude Pro $20 a month, or $17 billed yearly; Max $100; Ultra $200	Mac, Windows, Linux	Not on a Claude plan
Enjoy	Antigravity beside other labs’ agents, in an app	A monthly subscription, free for 7 days, with your phone and other computers (pricing)	Mac, Windows, Linux	Through Antigravity or OpenCode

Prices are per month as each maker lists them in the US, before tax. The open-source agents cost nothing themselves, and the model behind them may.

What happened to Gemini CLI

Gemini CLI still exists, but not for most of the people who used it. Google’s README now opens with “No free tier or Google AI plans,” and says individual access “moved to Antigravity CLI on June 18, 2026.” Its transition post from May 19 said Gemini CLI would “stop serving requests for Google AI Pro and Ultra, as well as those using it free of charge using Gemini Code Assist for individuals.”

Gemini CLI’s own quota page still lists 1,000 requests a day for a Google account, which can make it look as if nothing changed. The README and the transition post are the ones to believe.

Who keeps it: Google says organizations with a Gemini Code Assist Standard or Enterprise license keep their access, and that it “will remain accessible via paid Gemini and Gemini Enterprise Agent Platform API keys.”

A free Gemini API key still works elsewhere

Google’s sign-in went away, but the Gemini API’s free tier didn’t. Its pricing page lists Gemini 3.8 Flash as “Free of charge” on the free tier, and OpenCode, Cline, Qwen Code and Aider can use a key from Google AI Studio. There are two conditions:

• Google can use what you send. On the free tier, the same table answers “Used to improve our products” with “Yes.” On the paid tier it says “No.”
• The top model costs money. “Gemini 4 Argon is not available on the Free tier.” On the paid tier it’s $2 per million input tokens and $10 per million output through November 30, then $4 and $20. Gemini 3.8 Flash is $0.75 and $3.75 through December 31, then $1.50 and $7.50.

If you have old model names in a config, Gemini 2.5 Pro and Flash get rerouted on December 1.

Antigravity CLI

Best for: moving with the least work, on the same Google account.

Antigravity CLI, run as agy, is what Google moved everyone to. Its migration guide says it reads your GEMINI.md and AGENTS.md files unchanged, offers to convert your settings on first launch, and turns extensions into plugins with agy plugin import gemini. MCP servers move to their own mcp_config.json.

Price: $0 a month, with “Basic weekly rate limits.” Google AI Pro is $19.99 a month, and Ultra is $99.99 or $199.99. A free account gets Gemini 3.8 Flash, and Gemini 4 Argon starts with paid Pro. Claude Sonnet 4.6 and Opus 4.6 are on free accounts too, for now: “Will be removed on November 18, 2026,” says its models page, as our memo explains.

Watch for: “there won’t be 1:1 feature parity right out of the gate.” Free accounts get “A meaningful quota, refreshed weekly,” not Gemini CLI’s daily requests, and there’s no support for “Bring-your-own-key (BYOK).” Its GitHub repository holds a README, examples and a changelog, with no source code or license file, so you give up open source. For personal accounts, Google’s terms say “Google employees and contractors may access, view, review and use Interactions,” unless you change your preference in settings. agy -p is meant for scripts and CI, but wrapping it in third-party apps “is not supported under consumer plans,” says its FAQ. When the quota runs out, our quota guide has the options.

OpenCode

Best for: keeping an open-source terminal agent, and choosing the model.

OpenCode is “an open source agent that helps you write code in your terminal, IDE, or desktop,” under the MIT license. It says “Free models included or connect any model from any provider, including Claude, GPT, Gemini and more,” and you can log in with a ChatGPT Plus or Pro or GitHub Copilot account. Its downloads page has desktop apps for macOS, Windows and Linux.

Price: “OpenCode is 100% free to use.” Its optional Go plan is $10 a month and Go Plus is $40, both for open coding models. Other providers charge their own prices.

Watch for: its free models are for a limited time, and some may use your data. Its docs say of Big Pickle, “During its free period, collected data may be used to improve the model,” and that Flint 40B Free’s data “may be used to improve the model” and is “retained for 30 days.” Our OpenCode vs Claude Code guide has more.

Cline

Best for: your own Gemini key in VS Code or a terminal, with no subscription needed.

Cline calls itself “The open source coding agent in your IDE, terminal, & desktop,” under Apache 2.0. Install its CLI with npm i -g cline, or use the VS Code extension, and its Gemini setup takes a key from Google AI Studio.

Price: “Cline is free for individual developers. Pay only for AI inference on a usage basis.” ClinePass, for open coding models, is “a flat $9.99/month subscription.”

Watch for: its free models are “offered on a rotating, limited-time basis,” up to “a limited usage quota,” so they aren’t something to plan around. The JetBrains extension is on its Enterprise plan, and its desktop app is for macOS and Windows.

Qwen Code

Best for: something close to Gemini CLI’s roots, with many providers.

Qwen Code says it “was originally based on Google Gemini CLI v0.8.2,” though “Starting from Qwen Code v0.1, we stopped syncing with upstream.” It’s under Apache 2.0, installs on Linux, macOS and Windows, and “Supports OpenAI, Anthropic, Gemini, and Qwen APIs. Any third-party provider or local model.”

Price: the agent is free. You pay for the model: an API key, or Alibaba’s Coding Plan or Token Plan, whose prices are on Alibaba’s own pages.

Watch for: there’s no free sign-in anymore. “The Qwen OAuth free tier was discontinued on 2026-04-15,” its docs say.

Aider

Best for: pair programming in a terminal, where Git keeps every change.

Aider is “AI Pair Programming in Your Terminal,” under Apache 2.0, and “can connect to almost any LLM, including local models.” For Gemini, its docs say “You’ll need a Gemini API key,” and aider --list-models gemini/ shows what’s there.

Price: free, plus your model’s prices.

Watch for: it’s quiet. Its latest GitHub release is v0.86.0, from August 9, 2025, and its README still says it “works best with Claude 3.7 Sonnet, DeepSeek R1 & Chat V3, OpenAI o1, o3-mini & GPT-4o.” Its Gemini shortcut still points to Gemini 2.5 Pro, so name a current model yourself.

Codex CLI

Best for: moving to OpenAI’s models, and keeping an open-source terminal client.

Codex CLI is “a coding agent from OpenAI that runs locally on your computer,” under Apache 2.0, with installers for Mac, Linux and Windows.

Price: OpenAI’s pricing puts it on Plus, at $20 a month: “Codex on the web, in the CLI, in the IDE extension, and on iOS.” Pro starts at $100. An API key works too, for “Codex in the CLI, SDK, or IDE extension.”

Watch for: the free route isn’t in the terminal. Free and Go get “GPT-6 Luna at Standard speed in the desktop app, subject to rollout.” And on Plus, OpenAI’s estimates of messages per five hours “are not fixed message limits.” Our Codex alternatives guide compares more.

Claude Code

Best for: moving to Claude, on a Claude plan.

Claude Code is Anthropic’s coding agent, for the terminal, IDEs and its desktop app. It runs on your own Claude account, and it replaces Gemini rather than running it.

Price: Claude’s pricing has Pro at $20 a month, or $17 billed yearly; Max at $100; and Ultra at $200.

Watch for: there’s no free way in: “The free claude.ai plan does not include Claude Code access.” And “Pro and Max plans have a five-hour session limit and a weekly limit,” says Anthropic, “shared across Claude and Claude Code.” Our Claude Code vs Codex guide helps choose between the two.

Enjoy

Best for: keeping Gemini through Antigravity, with other labs’ agents a click away, in an app rather than a terminal.

Enjoy runs Antigravity on its free weekly quota or your Google AI Pro or Ultra plan, along with Claude Code, Codex, Grok Build, Muse Code, and OpenCode with the providers you set up in it, on Mac, Windows and Linux. What it adds:

• Switch agents mid-conversation. When Antigravity’s weekly quota runs out, or a job needs another lab’s model, hand it to another agent: “The new agent gets the conversation so far, including replies from other agents, before it works on your message.”
• Limits in view. Enjoy shows “the limits Google shows for each group of models.” When one runs out, the conversation waits for the reset and carries on, or you switch.
• Every project at once. Each conversation runs its own agent, and JOY keeps an eye on all your projects.

Price: Enjoy is one monthly subscription, and the first seven days are free, with a card to start. It includes your phone and any browser and your other computers, and each teammate you invite needs a subscription of their own. See pricing.

Watch for: it isn’t a free answer. Enjoy is a monthly subscription on top of your AI plans once the free week ends, and it isn’t open source, so if those are why you liked Gemini CLI, use OpenCode or Cline. It’s an app, not a terminal agent. Enjoy runs Antigravity through the server Google offers for other apps, not by wrapping agy, and in Enjoy it runs Gemini models only. Its default mode “approves every action, including commands and file changes,” and two other modes ask first. Enjoy doesn’t add one lab’s models to another’s account: Claude Code still needs a Claude plan, and Codex CLI a plan from Plus up. Scheduled workflows run only while your computer is awake with Enjoy open.

When to stay with Gemini CLI

Stay if your organization has a Gemini Code Assist Standard or Enterprise license, or Code Assist for GitHub on Google Cloud: Google says “your access remains unchanged.” Stay, too, if you run it on a paid API key and your scripts depend on it, though its README now calls those keys “legacy.” It’s still open source under Apache 2.0 and still ships, with v0.67.0 out on November 4, 2026.

Just know what Google says it’s for now: “Not optimized for Gemini 4+,” and “Not tuned for the Gemini 4 series.” If you want Gemini 4 Argon at its best, Google points you to Antigravity CLI.

If you signed in with a personal Google account or a Google AI plan, you can’t stay. Antigravity CLI is the shortest way forward, and everything above is for when its quota, its closed code or its models don’t suit you.

How we compared

We read each maker’s homepage, pricing, plans, docs and GitHub repository on November 5, 2026, and every one is linked below. We quote their own words for the limits, and we don’t count features a maker hasn’t documented. We don’t rank these by benchmark here: our best coding agent guide does that. Plans and models in this category change often, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.