How much does Gemini 4 Argon cost?

In the Gemini API (paid tier only), $2 per 1M input tokens and $10 per 1M output tokens through Nov 30, 2026, then $4 and $20 from Dec 1, 2026. Output includes thinking tokens.

Is Gemini 4 Argon free?

Not in the API: Google lists it as unavailable on the free tier. In the Gemini app it's included with Google AI Pro and Ultra subscriptions.

Can I use Gemini 4 Argon in GitHub Copilot?

Yes, on Copilot Pro+, Max, Business and Enterprise since Oct 13, 2026, billed at provider list pricing under usage-based billing.

Does Gemini CLI support Gemini 4 Argon?

Gemini CLI's README says it has not been optimized for Gemini 4 Argon and recommends Antigravity instead.

← All posts

Gemini · Pricing · Coding agents

Gemini 4 Argon pricing: every way to use it for coding (API, Copilot, Cursor, Devin Desktop)

Published October 20, 2026 · Facts checked against official vendor pages on October 20, 2026 · Rankings are independent of sponsorships

Google released Gemini 4 Argon on October 13, 2026, calling it its frontier model for long, multi-step work like large refactors and code migrations. It is already available in GitHub Copilot, Cursor and Devin Desktop, each with its own billing rules, and Google's API price doubles on December 1. Here is what it costs everywhere you can use it for coding today, read from each vendor's own pricing or docs page on October 20, 2026.

AGemini AGitHub Copilot SCursor ADevin Desktop (formerly Windsurf)AGemini CLI

The short version

Cheapest way to try it right now: Devin Desktop, which lists Argon at an extra 50% off its introductory price until October 27, 2026 ($1 input / $5 output per 1M tokens), billed against your plan's quota and extra usage.

If you already pay for an editor: GitHub Copilot (Pro+ and up) and Cursor both bill Argon on top of your plan at roughly Google's list price. It is never part of the usage included in a Cursor plan.

If you want it in a chat app: the Gemini app on Google AI Pro or Ultra. For Google's own agentic coding tool, Antigravity, you need Ultra.

Budget for the price change: every token-billed route below is on introductory pricing, and Google's standard API rates double from December 1, 2026.

Gemini API price (paid tier only)

Argon is not available on the Gemini API's free tier. On the paid tier, Google's pricing page lists these rates for the model ID gemini-4-argon:

Input: $2.00 per 1M tokens through Nov 30, 2026; $4.00 from Dec 1, 2026.
Output (including thinking tokens): $10.00 per 1M through Nov 30; $20.00 from Dec 1.
Context caching: $0.10 per 1M through Nov 30; $0.20 from Dec 1. Implicit caching only, explicit caching isn't supported.
Batch API: Google advertises a 50% cost reduction on the paid tier, and lists Argon batch rates of $1 input / $5 output per 1M through Nov 30.
Grounding with Google Search: 5,000 free requests per month shared across Gemini 3 and newer models, then $14 per 1,000.

What a coding task might cost (illustrative math)

Thinking tokens bill as output, so long reasoning gets expensive faster than prompt size suggests. As a worked example, a single agent run that reads 200,000 tokens of code and writes 20,000 tokens (thinking plus answer) costs 0.2 × $2 + 0.02 × $10 = $0.60 at the introductory rate, and $1.20 after December 1. Cached re-reads of the same files cost a twentieth of fresh input, so agents that keep a stable prefix get cheaper.

These numbers are our arithmetic on Google's published rates, not a benchmark. Real runs vary a lot with effort level and how many files the agent opens.

GitHub Copilot

GitHub made Argon generally available on October 13 for Copilot Pro+, Max, Business and Enterprise (Free and Pro aren't listed). It is billed at provider list pricing under Copilot's usage-based billing, and you can pick it in VS Code, Visual Studio, Copilot CLI, the Copilot coding agent, the Copilot app, JetBrains IDEs, Xcode and Eclipse. GitHub says rollout is gradual. Our GitHub Copilot page lists the current plans: Pro+ is $39/mo and Max $100/mo.

Cursor

Cursor's docs list Gemini 4 Argon with a 200k context window (1M max) and four effort levels: high, medium (default), low and minimal. It is a third-party model, so it isn't part of the usage included with Cursor plans: on The X Subscription it's billed from your prepaid credit at $2 input, $0.10 cache read and $10 output per 1M tokens, and on legacy Cursor plans it draws from the Other Models pool. See our Cursor page for how the bundle plans work, and our Cursor vs Devin Desktop comparison for the bigger picture on which models each plan covers.

Devin Desktop (formerly Windsurf)

Devin Desktop bills self-serve quota and extra usage by the token cost of the model you choose. Its models page lists Argon at introductory pricing of $2 input / $10 output / $0.10 cache input per 1M through December 1, 2026, with an additional 50% off ($1 / $5 / $0.05) through October 27, 2026, then regular pricing of $4 / $20 / $0.20. Enterprise ACU plans get 75% off list price until Oct 27 and 50% off until Dec 1. Details on plans are on our Devin Desktop page.

Gemini app, Antigravity and Android Studio

Google AI Pro and Ultra subscribers get Argon in the Gemini app, and the Gemini macOS app now includes Antigravity's agentic coding interface. Google's launch post puts Argon in Antigravity for Google AI Ultra subscribers; developers on the paid Gemini API can also use it in Google AI Studio and Android Studio. Current Google AI plan prices are on our Gemini page.

Gemini CLI users: read this first

Gemini CLI's own README now says it has not been optimized for Gemini 4 Argon and tells you to use Antigravity for it. It also notes that for unpaid-tier and Google One users, Gemini CLI was replaced by Antigravity CLI on June 18, 2026; Google's transition post says the CLI keeps working with paid Gemini API keys and Gemini Code Assist Standard or Enterprise licenses. We've updated our Gemini CLI page to reflect that it no longer has a free hosted quota for individuals.

Make Argon (or any model) work better on your repo

Google's launch post describes the ADK team giving Argon their .agents/skills rulebook plus a one-paragraph goal and letting it work overnight. The lesson carries to any model: clear, written project rules matter as much as the model. Start with our free AGENTS.md starter, a single file that agents such as Codex, GitHub Copilot's coding agent and Cursor can read, and keep our free vibe debug prompts handy for when an agent gets stuck in a fix loop. For Next.js projects, the Cursor Rules: Next.js 16 pack adds 12 scoped rules plus AGENTS.md and CLAUDE.md.

FAQ

How much does Gemini 4 Argon cost?: In the Gemini API (paid tier only), $2 per 1M input tokens and $10 per 1M output tokens through Nov 30, 2026, then $4 and $20 from Dec 1, 2026. Output includes thinking tokens.
Is Gemini 4 Argon free?: Not in the API: Google lists it as unavailable on the free tier. In the Gemini app it's included with Google AI Pro and Ultra subscriptions.
Can I use Gemini 4 Argon in GitHub Copilot?: Yes, on Copilot Pro+, Max, Business and Enterprise since Oct 13, 2026, billed at provider list pricing under usage-based billing.
Does Gemini CLI support Gemini 4 Argon?: Gemini CLI's README says it has not been optimized for Gemini 4 Argon and recommends Antigravity instead.

Free and paid rules for your agents

AGENTS.md starter for any stack (free)
Vibe debug prompts (free starter) (free)
Cursor Rules: Next.js 16 ($9): 12 scoped Cursor rules, AGENTS.md and CLAUDE.md for Next.js 16.4 with Cache Components.
Vibe Debug Pack ($9): 42 symptom prompts, a decision tree and git rescue to break AI fix loops.

Related changelog

Sources

Prices change often. Always confirm on the vendor's site before buying. See the full tier lists and the changelog.