Google makes Gemini 4 Argon generally available — paid-tier API in AI Studio and Android Studio on October 13, Pro and Ultra get it in the Gemini app, and per The Verge the intro price is $2/$10 per million tokens with a doubling coming at the end of November

4 sources

On October 13, 2026 (Tuesday, calendar-verified), Google made Gemini 4 Argon generally available. The Gemini API release notes carry the entry "Gemini 4 Argon (gemini-4-argon) generally available (GA)" — "our most capable model built for complex coding, agentic, math, and science" tasks. The developer post: "Starting today, developers can build with Gemini 4 Argon" — "Start building using the paid tier of the Gemini API in Google AI Studio and Android Studio" — "Google AI Ultra subscribers can try it in AI Studio, and our agent-first development platform, Google Antigravity". Framing: "our frontier model for complex workflows," for "long-horizon workflows that require deep reasoning and structural precision". The same day, the app post: "Gemini 4 Argon is now available in the Gemini app" — "Google AI Pro and Ultra subscribers can now access Gemini 4 Argon and give Gemini tasks to get more done." — "Try these updates starting today." Behavior claims: "answers questions more directly with less hedging"; "doesn't repeat itself more than necessary, and skips clichés throughout the responses"; on longer tasks "keeps you in the loop by sharing its progress more often and more effectively". Plus: Antigravity's agentic coding interface is now inside the Gemini macOS app; a chat-to-tasks toggle works while you're away; and Gemini connects to "more than 40 apps" including Instacart, OpenTable, Adobe, and Shopify. Per The Verge (Robert Hart, October 13): "The model is rolling out today to Google's higher-tier Pro and Ultra subscribers in the Gemini app and AI mode in Search," and "It is not clear when — or if — Argon will be available to Plus subscribers or free users." The Verge's pricing report: "The model is launching with an introductory price of $2 per million input tokens and $10 per million output tokens" — roughly doubling at the end of November when the introductory period finishes, per the outlet. Per The Verge, Google claims Argon beats GPT-6 Astra, Claude Fable 5.1, and Claude Opus 5.5 in coding, cybersecurity, science, finance, and law — a vendor claim with no published figures.

Why now

The model your audience has been waiting to build on is now generally available. Per the release notes, Gemini 4 Argon is "generally available (GA)" as "our most capable model built for complex coding, agentic, math, and science" tasks — and per the developer post the door in is concrete: "Start building using the paid tier of the Gemini API in Google AI Studio and Android Studio". Second layer, the consumer side is real but gated: per the app post "Google AI Pro and Ultra subscribers can now access Gemini 4 Argon and give Gemini tasks to get more done," and per The Verge "It is not clear when — or if — Argon will be available to Plus subscribers or free users." The tier map is the explainer. Third layer, the price is the story your audience will screenshot: per The Verge, "The model is launching with an introductory price of $2 per million input tokens and $10 per million output tokens" — with the outlet reporting those costs roughly double at the end of November, a deadline your viewers should see coming. Fourth layer, the benchmark fight is Google's own claim for now: per The Verge, Google says Argon beats GPT-6 Astra, Claude Fable 5.1, and Claude Opus 5.5 in coding, cybersecurity, science, finance, and law — with no published figures — so the credible angle is 'Google's claim, numbers pending', not a victory lap.

Why it is worth publishing

The most build-relevant card of the day: a frontier model hit GA across API, IDE, and app surfaces on the same day, and the price beat — an intro rate that roughly doubles at the end of November, per The Verge — gives your audience a concrete 'start now' reason. The differentiation play is the tier-and-price map: which surface takes which subscription, where the free tier is (nowhere yet, per The Verge's open question), and which numbers are Google's claims versus The Verge's reporting. Creators who publish the honest map own the launch beat; everyone else will screenshot the benchmark claim.

Evidence basis

Four opened sources: Google's developer post (page date label 'Oct 13, 2026', Tuesday calendar-verified; opened and read in full), the Gemini API release notes (October 13 GA entry; opened and read in full), Google's Gemini app post (page date label 'Oct 13, 2026'; opened and read in full), and The Verge (Robert Hart, byline Oct 13, 4:00 PM UTC, Tuesday calendar-verified; opened and read in full). Calendar checks: October 13, 2026 = Tuesday; October 15 (this card's publish date) = Thursday. Attribution discipline: the intro pricing ($2/$10, roughly doubling at the end of November) and the 'AI mode in Search' surface exist only in The Verge's report — the changelog defers pricing to a separate guide and no official pricing page was opened this run, so pricing never ships as Google's official or permanent pricing. Benchmark claims stay Google-claims-per-The-Verge with no figures. Free-tier availability stays an open question. Announcement-voice 'starting today' phrases are anchored to October 13 and never repeat as the card's own voice. Search-summary discipline: a 'frontier 5.5' benchmark phrase visible in search summaries was checked against the opened sources, not found, and never written.

“Google's frontier model Gemini 4 Argon is now generally available — API builders can start on the paid tier, and per The Verge its intro price is roughly one-fifth the cost of GPT-6 Astra.”

Angle

Frame it as 'the frontier model just opened its doors — here's the honest map' in three beats. Beat one, where it actually is: per the release notes it's "generally available (GA)" as gemini-4-argon; per the developer post the build doors are "the paid tier of the Gemini API in Google AI Studio and Android Studio" plus AI Ultra in AI Studio and Antigravity; per the app post Pro and Ultra subscribers get it in the Gemini app. Beat two, the price with its clock: per The Verge, "The model is launching with an introductory price of $2 per million input tokens and $10 per million output tokens" — roughly doubling at the end of November, per the outlet — show the deadline on screen. Beat three, the claim discipline: per The Verge, Google says Argon beats GPT-6 Astra, Claude Fable 5.1, and Claude Opus 5.5 in coding, cybersecurity, science, finance, and law — with no published figures — so run your own small task comparison on camera and label Google's claim as Google's claim.

Format

Long-form explainer

Demo idea

A 'tier and price map' walkthrough: one screen showing where Argon lives (paid API in AI Studio and Android Studio, Ultra in AI Studio and Antigravity, Pro and Ultra in the Gemini app, Search AI mode per The Verge), then the intro $2/$10 pricing with the end-of-November doubling deadline per The Verge, then a live build demo — one small agentic task in AI Studio — with on-screen labels separating Google's claims from The Verge's report.

Platform notes

The pricing ($2/$10 intro, roughly doubling at the end of November) is The Verge's report — never Google's official pricing, never permanent; 'introductory' stays attached. The benchmark comparison is Google's claim via The Verge, no figures — never stated as independent fact or charted with added numbers. Surface scope stays split per source: paid-tier API in AI Studio and Android Studio, AI Ultra in AI Studio and Antigravity, Google AI Pro and Ultra in the Gemini app, AI mode in Search attributed to The Verge. Free tier: "It is not clear when — or if — Argon will be available to Plus subscribers or free users." (The Verge) — never 'available to everyone'. 'Starting today' is October 13 announcement voice — the card says October 13. our most capable model' keeps Google's own scope — never 'the world's most capable model'.

Usable claims

On October 13, 2026 (Tuesday, calendar-verified), Google made Gemini 4 Argon generally available. The Gemini API release notes carry the entry "Gemini 4 Argon (gemini-4-argon) generally available (GA)", describing it as "our most capable model built for complex coding, agentic, math, and science" tasks. Google's developer post announced: "Starting today, developers can build with Gemini 4 Argon" — "Start building using the paid tier of the Gemini API in Google AI Studio and Android Studio" — and "Google AI Ultra subscribers can try it in AI Studio, and our agent-first development platform, Google Antigravity". The post describes "our frontier model for complex workflows," aimed at "long-horizon workflows that require deep reasoning and structural precision". The same day, Google's app post said: "Gemini 4 Argon is now available in the Gemini app" — "Google AI Pro and Ultra subscribers can now access Gemini 4 Argon and give Gemini tasks to get more done." — and "Try these updates starting today." Per the app post, Argon "answers questions more directly with less hedging," "doesn't repeat itself more than necessary, and skips clichés throughout the responses," and on longer tasks "keeps you in the loop by sharing its progress more often and more effectively"; the post also puts Antigravity's agentic coding interface inside the Gemini macOS app, adds a chat-to-tasks toggle for working on tasks while away from the computer, and connects Gemini to "more than 40 apps" including Instacart, OpenTable, Adobe, and Shopify. Per The Verge (Robert Hart, 4:00 PM UTC, October 13): "The model is rolling out today to Google's higher-tier Pro and Ultra subscribers in the Gemini app and AI mode in Search," and "It is not clear when — or if — Argon will be available to Plus subscribers or free users." The Verge reports the launch pricing: "The model is launching with an introductory price of $2 per million input tokens and $10 per million output tokens" — introductory, with costs roughly doubling at the end of November when that period finishes, per the outlet. Per The Verge, Google claims Argon beats GPT-6 Astra, Claude Fable 5.1, and Claude Opus 5.5 in coding, cybersecurity, science, finance, and law — a vendor claim the outlet relays without published figures — and relays Google's framing that Argon delivers "frontier reasoning and improved performance across complex workflows". The outlet's background: the GA follows a teaser "a few weeks ago" and a limited release to "trusted cyber defenders"; internally Google has used the model for patching cybersecurity vulnerabilities, halving compute for YouTube comments, and debugging Wing's drone physics simulator, per the article.

Evidence pipeline

From the news

Breakdown

A model-launch story where pricing attribution and surface scope are the content. Layer one, the stable facts: per the release notes, "Gemini 4 Argon (gemini-4-argon) generally available (GA)" on October 13, described as "our most capable model built for complex coding, agentic, math, and science" tasks — Google's own scope, never flattened to 'the world's most capable model'. Layer two, the price wears a reporter's label: "The model is launching with an introductory price of $2 per million input tokens and $10 per million output tokens" is The Verge's report — the changelog defers pricing to a separate guide, no official pricing page was opened, and the intro rate roughly doubles at the end of November, per the outlet — so 'introductory' stays attached and Google-official framing never forms. Layer three, the benchmark claims stay Google-claims-per-The-Verge: beating GPT-6 Astra, Claude Fable 5.1, and Claude Opus 5.5 across coding, cybersecurity, science, finance, and law ships with no figures attached and never as independent fact; a 'frontier 5.5' phrase visible in search summaries was checked against the opened sources, not found, and never written. Layer four, the tier map is the split that holds: paid-tier API in AI Studio and Android Studio, AI Ultra in AI Studio and Antigravity, Google AI Pro and Ultra in the Gemini app — each from its own official post — with 'AI mode in Search' attributed to The Verge, and the free tier left as the open question The Verge states ("It is not clear when — or if — Argon will be available to Plus subscribers or free users."). Layer five, announcement voice: 'Starting today' and 'Try these updates starting today' are October 13 phrases inside attributed quotes; the card's own voice says October 13. Editor's rules: reporter-labeled prices, vendor-labeled benchmarks, and a tier map per source.

Sources

Risks

Before publishing, re-check each layer: the pricing stays attributed to The Verge with 'introductory' attached and the end-of-November increase intact; the benchmark claims stay Google-claims-per-The-Verge with no figures added; the surfaces stay split per source (paid API in AI Studio/Android Studio, Ultra in AI Studio/Antigravity, Pro and Ultra in the Gemini app, Search AI mode attributed to The Verge); free-tier stays an open question; 'starting today' stays anchored to October 13; 'most capable' keeps Google's own scope; and the GA gap stays the outlet's framing. If your script compresses any of these, cut the detail rather than round it.

Demo ideas

Tier-and-price map: where Argon lives per surface, the intro $2/$10 pricing with the end-of-November doubling per The Verge, and a live small agentic task in AI Studio
Claims-versus-numbers card: Google's benchmark claim (GPT-6 Astra, Claude Fable 5.1, Claude Opus 5.5 across five domains, per The Verge) on screen with 'no published figures' labeled, followed by your own one-task comparison