The Gemini app is expanding access to its Gemini 4 Argon model, rolling out the latest AI enhancements to both AI Pro and AI Ultra users. Originally limited to the highest tier, the update introduces performance, coding, and interface improvements designed to streamline workflows and boost efficiency across Google’s AI ecosystem.
—
## Broadening Access: Pro Users Join the Argon Club
Until recently, only AI Ultra subscribers—whose tier starts at $99.99 per month—could use Gemini 4 Argon. As of October 2026, Google has opened access to AI Pro subscribers as well. The upgrade promises “frontier reasoning and improved performance across complex workflows,” along with more precise, less hedged answers that avoid clichés and unnecessary repetition. One of the biggest changes: Argon now follows style instructions more reliably than earlier Gemini models.
—
## Smoother for Long Tasks and Better Coding Performance
Argon offers consistent status updates during longer tasks, helping users stay informed about the AI’s progress. For coding work, Gemini app on macOS users will find a new “agentic coding interface” through Google Antigravity—enabling the creation of apps and web tools with more agility.
Other notable UI changes include a toggle in the main chat window to shift between general chat mode and task-oriented work (which was previously called “Spark”).
—
## More Models in the Picker, Argon in AI Mode, Expanded Tool Support
AI Pro and AI Ultra users can now choose from multiple models in Gemini: 3.5 Flash-Lite, 3.8 Flash, 3.1 Pro, and the new 4 Argon. Additionally, the new model is available in AI Mode as well, giving users broader context-aware capabilities.
The Argon model requires the AI Ultra plan when used with:
– Google Antigravity (desktop app and CLI version), which already uses Argon internally across projects like Chrome, Gemma, and YouTube.
– AI Studio and the Canary channel of Android Studio.
—
## Real-World Gains: Speed and System Optimization
Internally, Google saw substantial gains using Argon in production environments:
– The Chrome team separated the core web‐page handling from a legacy codebase of about 6 million lines, bypassing the need to rewrite vast sections of Chromium and enabling smoother tab and panel transitions without reloading pages—improving stability.
– Using Argon in an implementation with vLLM on A100 GPUs, Google achieved a 2.85× increase in end-to-end inference speed—from ~37,900 to over 108,000 tokens per second—maintaining 100% token parity with zero quality loss.
– Argon helped reduce YouTube’s comments frontend CPU workload by roughly half; trimming server overhead from redundant emoji rendering alone (around 45%) brought down server latency by about 30 ms and end-to-end client latency by 20 ms.
—
## Pricing and Special Offers
Argon is now available through the Gemini API with an introductory pricing model:
– **Until November 30, 2026:** $2 per million input tokens and $10 per million output tokens.
– **After November 30, 2026:** Rates increase to $4 per 1M input tokens and $20 per 1M output tokens.
Google is also granting free access to its highest usage tier in Antigravity for six months to 10,000 scientists who apply.
—
## What This Means for Users
For AI Pro members, Argon unlocks higher performance and smarter interactions. Ultra users see the benefits already amplified across Google tools like Antigravity, Chrome, Gemma, and YouTube. All tiers now include capabilities such as model switching—including the new Argon model, better task progress visibility, and improved style adherence.
With improved coding workflows, faster inference speeds, and substantial resource optimizations, Google’s update appears aimed at strengthening productivity for developers, scientific users, and creators alike.
—
This article is AI-generated content. Please verify the information independently before taking any action based on this article. before returning it.
