---
title: "Nothing is free: Google rolls out flagship Gemini 4 AI"
description: "Google has started the global rollout of its flagship Gemini 4 Argon AI model with advanced coding features. The novelty is available via subscriptions starting at $99 per month and through API with temporary discounts until late November."
date: 2026-10-14T08:14:00.000Z
lang: en
url: https://xab.info/en/posts/google-rolls-out-flagship-gemini-4-argon
tags: [google, gemini-4, artificial-intelligence, argon, api, tech-news]
publisher: "XAB.info"
---

# Nothing is free: Google rolls out flagship Gemini 4 AI

![Interface of the flagship Google Gemini 4 Argon AI model](https://xab.info/media/2026/10/14/google-razvorachivaet-flagmanskiy-ii-gemini-4-argon/google-razvorachivaet-flagmanskiy-ii-gemini-4-argon-1.webp)

## 🎯 Key Points

- Google launched the flagship Gemini 4 Argon AI model with enhanced reasoning capabilities.
- The model is available in the Google ecosystem, in Antigravity and AI Studio products on the AI Ultra tier starting at $99 per month.
- Temporary API rates are $2 per million input and $10 per million output tokens with a cache discount until November 30.

Google has officially begun the massive rollout of its new flagship artificial intelligence model, named Gemini 4 Argon. The presentation of this advanced development has drawn immense attention from the IT community, as the corporation positions the system as a qualitative leap in generative intelligence, reasoning, and code writing. The launch of the technology comes against the backdrop of fierce competition in the global AI market, where each new iteration must prove its superiority not just in words, but in practice.

### Technological Features and Capabilities of Argon

The characteristics of the top-tier AI remain unchanged, yet developers have focused heavily on deepened logic and safety. Google references recent benchmarks as proof of advanced reasoning and code agent capabilities. Additionally, the model is trained to actively reject malicious instructions, and a special built-in monitor can instantly halt processing if it detects an unpredictable or potentially dangerous workflow. Such measures are designed to minimize the risks of autonomous task execution.

The tech giant has already cited several striking examples of successful internal use of Argon within the corporation. Specifically, while optimizing YouTube comments, the neural network discovered redundant emoji rendering, the elimination of which reduced server load by an impressive 45 percent. In the Wing project drone physics simulator, the artificial intelligence found critical bugs that had remained in the code since 2018. The model also demonstrated excellent results in robotics, showing a noticeable step forward in spatial reasoning for three-dimensional environments.

### Contradictory Data

Despite impressive official demonstrations and internal corporate tests, serious controversies have flared up around the release of Gemini 4 Argon. Some specialized publications and independent experts criticized the novelty in the very first days after launch, claiming, with reference to their own tests, that the model allegedly demonstrates inflated benchmark results due to task-specific optimization and sometimes stumbles when writing complex production code. Google representatives are currently refraining from detailed comments regarding the criticism, insisting on the objectivity of standard metrics.

### Access Conditions and Pricing Policy

The new model is distributed on a commercial basis, confirming the corporation's concept of the high value of advanced computing. In the familiar Gemini consumer app, the model becomes available under the AI Pro subscription for $20 per month and AI Ultra, starting from $99 per month. In specialized developer and coding products such as Antigravity and AI Studio, Gemini 4 Argon is integrated exclusively on the most expensive AI Ultra tier due to massive computational costs.

For less complex and routine tasks, the company wisely retains the previous version, Gemini 3.1 Pro. For developers working via the paid API, the temporary token cost is set at $2 per million input tokens and $10 per million output tokens. Furthermore, a special 95 percent discount on cached tokens applies until November 30. After the promotion ends, standard rates will double—to $4 for incoming and $20 for outgoing tokens.

## 🔍 Fact-Check Verification

- [Бесплатно не работает: Google разворачивает флагманский ИИ Gemini 4](https://www.rbc.ua/ukr/news/bezkoshtovno-pratsyue-google-rozgortae-flagmanskiy-1791962949.html) - Основной источник информации о релизе модели и ценовой политике.
- [Google запускает бюджетную версию Gemini 2.5 Flash в Vertex AI и готовит локальное развертывание моделей](https://habr.com/ru/articles/899228/) - Контекст линейки моделей Google.
- [Google Gemini 4 выйдет до конца 2026 года](https://rozetked.me/news/48513-google-gemini-4-vyydet-do-konca-2026-goda) - Временные рамки и ожидания релиза.
- [Новую ИИ-модель Google Gemini 4 Argon раскритиковали в первый же день: модель мухлюет на бенчмарках и ...](https://www.ixbt.com/news/2026/10/03/440000-novuiu-ii-model-google-gemini-4-argon-raskritikovali-v-pervyi-ze-den-model-muxliuet-na-bencmarkax-i-buksuet-v-kode.html) - Источник критики модели и замечаний по бенчмаркам.

## ❓ FAQ

### Q: How much does a subscription to access Gemini 4 Argon cost?
**A:** In the consumer Gemini app, the model is available on AI Pro ($20/mo) and AI Ultra (from $99/mo) subscriptions. In Antigravity and AI Studio development environments, it is available exclusively on the AI Ultra tier.

### Q: What promotional tariffs apply to developers via the API?
**A:** Until November 30, a temporary token price applies: $2 per million input and $10 per million output tokens with a 95% discount on cached tokens. After November 30, prices will double.