# Gemini 4 Argon Hits AI Mode in 19 Countries [2026] > Source: > Published: 2026-10-19 09:55:36+00:00 Google has started rolling out [**Gemini 4 Argon**](https://deepmind.google/models/gemini/) inside **AI Mode**, the conversational search experience built into Google Search, giving **Google AI Pro** and **Google AI Ultra** subscribers a new model option for complex, multi-step queries. The expansion, confirmed by Google on October 13, 2026, and reported the same day by outlets including Search Engine Journal, SiliconANGLE, SE Roundtable, Yahoo Tech and The AI Insider, marks the first time Argon has reached ordinary consumer search traffic rather than developer tools or early-access partners. The move is part of a broader push by Google to put its newest flagship model in front of as many paying subscribers as possible, following a staged rollout that began with vetted cybersecurity defenders and worked outward from there. For a company trying to answer competitive pressure from [OpenAI’s ChatGPT](https://shattered.io/chatgpt-gemini-2-billion-visits-market-share-2026/) and Anthropic’s Claude lineup, putting a frontier model directly into the search bar that billions of people already use is one of the fastest ways to get it in front of real users. The move is the latest entry in this site’s ongoing [AI and machine learning coverage](https://shattered.io/ai-machine-learning/) tracking how fast Google, OpenAI and Anthropic are shipping new models. Don't miss new tech stories on Google Add Shattered once in the Google app and our stories show up in your news suggestions. [Add now](https://www.google.com/preferences/source?hl=en&q=https%3A%2F%2Fshattered.io%2F) ## Gemini 4 Argon Expands Into Google Search’s AI Mode Google’s own announcement, published to its blog on October 13, 2026, describes Gemini 4 Argon as a model built to tackle difficult, multi-step problems. According to Google, Argon is now available across several surfaces simultaneously: the Gemini app, the Gemini API through AI Studio, Gemini for business, Google Antigravity, and **AI Mode in Google Search**. SE Roundtable reported that Google executives Nick Fox and Robby Stein posted on X that “Gemini 4 Argon is now rolling out globally in Search!” alongside the broader announcement, framing the Search integration as a global rollout rather than a limited regional test. That framing matters because AI Mode has become Google’s primary vehicle for layering generative AI on top of traditional search results. Rather than a single answer box, AI Mode lets users have an extended, conversational exchange with a model that can pull in live web results, reason across multiple steps, and follow up on earlier parts of the conversation. Slotting Argon into that experience gives Pro and Ultra subscribers access to what Google is positioning as its most capable model yet, directly inside the search product hundreds of millions of people already open every day. Search Engine Journal, which first flagged the AI Mode update, noted that Google has been iterating on the model menu inside AI Mode for months, and that Argon is simply the latest addition to a lineup users can toggle between depending on the task at hand. ## What Changed Inside AI Mode’s “Ask Anything” Bar Mechanically, the change is straightforward. Search Engine Journal reported that users can now select Argon from the “+” icon inside AI Mode’s “Ask anything” input bar, the same menu location Google has used to introduce other Gemini models to Search over the past few months. Selecting Argon swaps the underlying model for that session, meaning a query that might otherwise be handled by a faster, lighter Gemini model gets routed instead to Argon’s larger reasoning capacity. Google has said Argon brings improvements to nuanced, multi-step tasks, intent understanding, instruction following, and extended back-and-forth conversations, according to reporting from Search Engine Journal and The AI Insider. In practice, that is aimed at queries that go beyond a single lookup: multi-part research questions, comparisons that require synthesizing several sources, or follow-up questions that build on context from earlier in the same AI Mode session. The AI Insider’s coverage also noted that Google is positioning Argon for complex professional workflows beyond casual search, including software engineering tasks, legal and financial knowledge work, and cybersecurity defense use cases. That positioning lines up with the model’s staged rollout, which gave cybersecurity teams early access before the broader Pro and Ultra subscriber base got it in AI Mode and the Gemini app. ## Who Qualifies for Gemini 4 Argon in AI Mode Access to Gemini 4 Argon inside AI Mode is gated behind a specific set of eligibility rules, according to Google’s own support documentation. Users must be at least 18 years old, signed in with a personal Google Account rather than a managed workspace or school account, and physically located in a supported country or territory. On top of the eligibility requirements, Google confirms that Argon usage in AI Mode is subject to daily usage limits, though the company’s public documentation does not disclose the exact figures for either the Pro or Ultra tier. That combination of gating rules is consistent with how Google has rolled out other high-compute AI features in Search before: age and account verification to keep the feature within supported consumer accounts, regional restrictions tied to where the underlying infrastructure and subscription billing are available, and usage caps to manage the computational cost of running a frontier-scale model on live search traffic at global volume. ## The Countries and Territories Where Argon Is Live Google’s supported-locations list for Argon in AI Mode, as published in its own help documentation, names the United States, Canada, Brazil, Mexico, Argentina, Chile, Colombia, Costa Rica, the Dominican Republic, Ecuador, Guatemala, Haiti, Jamaica, Nicaragua, Panama, Puerto Rico, Trinidad and Tobago, Uruguay, and Venezuela, along with several additional Caribbean territories. That is nineteen individually named countries and territories plus an unspecified number of smaller Caribbean markets bundled into the same support note. What stands out in that list is its geographic shape: it is almost entirely North, Central and South American, with heavy representation across the Caribbean. Google’s own documentation does not claim the list is exhaustive or permanent, and the company has a track record of widening country support for AI Mode features in stages rather than all at once. Readers outside the listed countries and territories should not assume Argon is unavailable to them going forward, only that it has not yet been confirmed for their region as of this rollout. ## Daily Usage Limits: Confirmed Caps and Open Questions Google confirms that daily usage limits apply to Gemini 4 Argon inside AI Mode, but it has not published the specific number of queries either Pro or Ultra subscribers get per day. SiliconANGLE’s reporting on the rollout noted that Ultra subscribers receive higher Argon usage limits than Pro subscribers, consistent with Ultra’s positioning as the more expensive, higher-ceiling tier across Google’s AI subscription lineup. Beyond that relative comparison, however, there is no publicly confirmed, current figure for exactly how many Argon queries either tier allows in AI Mode. That gap is worth flagging directly: multiple outlets covering the October 13 rollout, including SiliconANGLE and Search Engine Journal, describe the existence of daily caps without specifying their size. Until Google publishes exact numbers or a third party reliably benchmarks the limit through hands-on testing, any specific daily-query figure circulating online should be treated as unverified. ## Google AI Pro vs Google AI Ultra: Price and Access Gap Both subscription tiers unlock Argon in AI Mode, but the surrounding offer differs in price and depth. Yahoo Tech and SiliconANGLE’s coverage of the rollout put Google AI Pro at roughly $19.99 per month in the United States, with Google AI Ultra starting at roughly $249.99 per month. The roughly twelve-times price gap between the two tiers buys subscribers higher Argon usage limits, along with access to additional surfaces: Ultra subscribers get Argon in Google Antigravity, Google’s agentic development platform, while that access is not confirmed for Pro subscribers. | Plan | Approx. US Price | Argon in AI Mode | Argon in Gemini App | Argon in Antigravity | |---|---|---|---|---| | Google AI Pro | ~$19.99/month | Yes | Yes | Not confirmed | | Google AI Ultra | ~$249.99/month | Yes (higher daily limit) | Yes | Yes | | Gemini API / AI Studio | Pay-per-token | N/A | N/A | N/A | | Vetted cybersecurity defenders (early access) | Program-based | No (preceded public rollout) | No | No | That pricing structure places Google’s subscription ladder firmly in the premium tier of consumer AI products, a pattern this site has tracked across Google’s recent Gemini launches, including the [pricing and benchmark split](https://shattered.io/gemini-4-argon-pricing-benchmark-split-2026/) that accompanied Argon’s original API launch. ## Inside the Model: a 1-Million-Token Context Window Separate from the AI Mode consumer rollout, [Google’s developer-facing documentation](https://ai.google.dev/gemini-api/docs/latest-model) describes Gemini 4 Argon’s Gemini API as offering a unified input-output context window of up to **1 million tokens**, aimed at longer scientific, coding and agentic workloads. That scale of context window lets the model hold far more surrounding material in view at once than earlier, smaller-context Gemini releases, a jump this site previously covered when Argon’s token ceiling first expanded sixteen-fold in its [global API launch](https://shattered.io/gemini-4-argon-global-launch-16x-tokens-2026/). Google set the introductory API price for Argon at $2 per million input tokens and $10 per million output tokens, according to both Google’s own blog post and U.S. News’ reporting on the launch. Cached input tokens carry a 95% discount during the introductory period, a cost-cutting mechanism that favors workloads reusing the same long context repeatedly. Yahoo Tech reported that this introductory pricing window is scheduled to run only through November 30, 2026, after which the listed prices are expected to double. That timeline gives developers roughly six weeks from the October 13 announcement to lock in the lower rate before costs rise. None of the AI Mode consumer rollout changes that developer-side pricing directly, since AI Mode usage is bundled into the flat monthly Pro or Ultra subscription rather than billed per token. But the two tracks are connected: the same underlying model now has a consumer-facing entry point in Search and a metered, pay-per-token entry point for developers building on top of it through the API. ## From Vetted Cybersecurity Defenders to Everyday Search SiliconANGLE’s reporting on the October 13 rollout noted that Gemini 4 Argon reached Gemini app, Pro and Ultra subscribers roughly two weeks after the model first went to a narrower group: vetted cybersecurity defenders given early access to test the model’s capabilities in security-specific workflows. That staged approach, early access for a specialized professional group followed by a broader consumer rollout, extends to [Argon’s availability on Google Cloud](https://cloud.google.com/blog/products/ai-machine-learning/gemini-4-argon-available-on-google-cloud) for enterprise customers, and mirrors how Google and other major AI labs have increasingly handled the launch of high-capability models, using a controlled first phase to surface issues before scaling to general availability. The two-week gap between the cybersecurity-focused early access and the AI Mode consumer launch is short by historical AI rollout standards, suggesting Google moved quickly once the initial testing phase wrapped. It also means Argon has now completed a full arc, from specialized early access, to developer API availability, to the Gemini app and AI Mode, inside roughly a month of its original announcement. ## A Rapid-Fire Model Cadence: 3.7 Flash, 3.8 Flash, Now Argon Argon’s addition to AI Mode’s model menu is not an isolated event. Search Engine Journal’s coverage points out that the same “+” menu in AI Mode’s “Ask anything” bar previously added Gemini 3.7 Flash in August 2026 and [Gemini 3.8 Flash](https://shattered.io/gemini-3-8-flash-googl-stock-0-7-percent-2026/) in September 2026, the same model whose benchmark performance [this site covered](https://shattered.io/gemini-4-argon-pricing-benchmark-split-2026/) when it briefly lifted Alphabet’s stock. That puts Argon as the third new model option added to AI Mode in three consecutive months. | Month (2026) | Model Added to AI Mode | Source | |---|---|---| | August | Gemini 3.7 Flash | Search Engine Journal | | September | Gemini 3.8 Flash | Search Engine Journal | | October 13 | Gemini 4 Argon | Search Engine Journal, Google, SiliconANGLE | That monthly cadence signals Google is treating AI Mode’s model menu as a living product surface rather than a static feature, continuously swapping in newer Gemini releases as they clear internal testing. It also means Pro and Ultra subscribers have had to keep track of an increasingly crowded model picker, each addition carrying different strengths, context limits and intended use cases. ## Why Google Is Racing to Regain Ground on OpenAI and Anthropic Reuters’ reporting on Argon’s October 13 launch, carried by U.S. News, framed the release as part of Google’s effort to regain ground against rivals OpenAI and Anthropic following months of delays and a leadership shakeup at its DeepMind lab. That context matters for understanding why Google is pushing Argon into so many surfaces simultaneously, the API, the Gemini app, Antigravity, Gemini for business, and now AI Mode in Search, rather than staggering the rollout over a longer period. Putting a frontier model directly inside Search gives Google a distribution advantage that OpenAI and Anthropic do not have in the same form: a search product already used by billions of people, where a model upgrade reaches existing subscribers without requiring them to switch apps or change habits. That distribution edge has been central to Google’s AI strategy since it first layered generative features onto Search, and it is the same dynamic behind the [AI Overviews feature now appearing on close to 13% of US searches](https://shattered.io/google-ai-overviews-search-rules-2026/), according to this site’s prior coverage of that rollout. ## Competitive Landscape: AI Mode vs ChatGPT Search and Perplexity Google is not the only company fighting for share of AI-assisted search queries. OpenAI’s ChatGPT has its own search capabilities, and Perplexity has built its entire product around AI-native search and synthesis. Reporting gathered on Argon’s AI Mode rollout did not surface specific, current benchmark comparisons between Argon’s search performance and either rival, so any head-to-head performance claim beyond what Google itself has published should be treated as unconfirmed for now. What is measurable is traffic. This site’s earlier coverage of the broader AI search market noted that Gemini’s consumer surfaces had topped 2 billion visits as ChatGPT’s own traffic growth cooled, a sign that Google’s built-in distribution through Search and the Gemini app is translating into real usage volume rather than just announcement headlines. Layering Argon into AI Mode is a bid to keep that momentum going at a moment when OpenAI and Anthropic are both shipping aggressively on their own release schedules. Google’s own comparative benchmark releases for Argon have been mixed rather than uniformly dominant. This site previously reported that Argon won only a portion of its head-to-head AI benchmark tests against rival models at launch, a reminder that raw capability claims and real-world search usefulness do not always move in lockstep. ## Market Impact: What This Means for Search Behavior and Subscriptions For everyday Search users, the most immediate effect of this rollout is choice: AI Mode now offers multiple Gemini model options, with Argon positioned as the heavyweight pick for complex, multi-step questions. That could shift some search sessions that previously bounced between separate queries into longer, single AI Mode conversations, particularly for comparison shopping, research-heavy questions, or technical troubleshooting that benefits from a model that can track context across several turns. For Google’s subscription business, gating Argon’s AI Mode access behind Pro and Ultra rather than offering it to free-tier Search users is a direct incentive to upgrade. Combined with the higher usage limits on Ultra, the tiering creates a familiar freemium funnel: free users see AI Mode with more limited models, Pro subscribers get Argon with standard caps, and Ultra subscribers get Argon with a higher ceiling plus access to Antigravity. Whether that structure meaningfully increases Pro or Ultra signups is not yet measurable from the data available at launch, but it follows the same playbook Google has used with prior premium AI feature rollouts. ## Historical Context: How AI Mode Got Here AI Mode itself has gone through its own rapid evolution since launch, expanding from a limited experiment into a core part of Google Search’s interface for Pro and Ultra subscribers. The monthly addition of new Gemini models, 3.7 Flash, then 3.8 Flash, now Argon, reflects a broader pattern across Google’s AI division in 2026: ship frequently, iterate the model menu rather than waiting for a single large version bump, and use Search as the fastest distribution channel for whatever model is newest. That approach stands in contrast to the slower, more consolidated release cadence Google followed with earlier Gemini generations, where months could pass between meaningful consumer-facing model updates. The cybersecurity-first early access phase that preceded Argon’s public rollout also reflects a pattern seen elsewhere in the industry: giving specialized professional users first access to a new frontier model, both to stress-test real-world use cases and to generate credibility with a technically sophisticated audience, before expanding to general consumer availability. ## Predictions: Where Gemini 4 Argon and AI Mode Go Next - Google is likely to publish or leak exact daily usage limits for Argon in AI Mode as third-party testers and reviewers probe the caps in the weeks following the October 13 rollout. - Expect the supported-country list to widen beyond its current North, Central and South American and Caribbean focus, following the pattern of staged geographic expansion Google has used with past AI Mode features. - The introductory API pricing of $2/$10 per million tokens is scheduled to expire November 30, 2026, which should push some cost-sensitive developers to lock in usage before prices are expected to double. - Given the monthly cadence of 3.7 Flash, 3.8 Flash and Argon, another new model addition to AI Mode’s model menu before the end of 2026 would be consistent with Google’s recent release rhythm. - Competitive responses from OpenAI’s ChatGPT Search and Perplexity are likely in the coming weeks, even though specific comparative benchmarks against Argon in AI Mode are not yet available. These are forward-looking assessments based on Google’s current rollout pattern and published pricing schedule, not confirmed company plans, and should be read as analysis rather than fact. ## Frequently Asked Questions **What is Gemini 4 Argon?** Gemini 4 Argon is Google’s flagship AI model, announced October 13, 2026, and described by Google as built to handle difficult, multi-step problems. It is available through the Gemini API, AI Studio, the Gemini app, Gemini for business, Google Antigravity, and now AI Mode in Google Search. **Who can use Gemini 4 Argon in AI Mode?** Google requires users to be 18 or older, signed in with a personal Google Account, and located in a supported country or territory. Access is limited to Google AI Pro and Google AI Ultra subscribers. **How much does Google AI Pro or Google AI Ultra cost?** Reporting from Yahoo Tech and SiliconANGLE puts Google AI Pro at roughly $19.99 per month in the United States, with Google AI Ultra starting at roughly $99.99 per month. **Is there a daily limit on Gemini 4 Argon queries in AI Mode?** Yes. Google confirms daily usage limits apply, and SiliconANGLE reported that Ultra subscribers get a higher limit than Pro subscribers, but Google has not published the exact query caps for either tier. **Which countries currently support Gemini 4 Argon in AI Mode?** Google’s support documentation lists the United States, Canada, Brazil, Mexico, Argentina, Chile, Colombia, Costa Rica, the Dominican Republic, Ecuador, Guatemala, Haiti, Jamaica, Nicaragua, Panama, Puerto Rico, Trinidad and Tobago, Uruguay, Venezuela, and several additional Caribbean territories. **How do I select Argon inside AI Mode?** According to Search Engine Journal, users can pick Argon from the “+” icon in AI Mode’s “Ask anything” bar, the same menu used to add earlier models like Gemini 3.7 Flash and 3.8 Flash. **What is the Gemini API pricing for Gemini 4 Argon?** Google’s introductory pricing is $2 per million input tokens and $10 per million output tokens, with a 95% discount on cached input tokens. That introductory pricing is scheduled to run through November 30, 2026, after which prices are expected to double, according to Yahoo Tech. **Did Gemini 4 Argon launch directly into AI Mode, or was there an earlier access phase?** SiliconANGLE reported that Argon first went to vetted cybersecurity defenders roughly two weeks before reaching Gemini app, Pro and Ultra subscribers and AI Mode in Search on October 13, 2026.