Gemini 4 Argon: Built to Solve Your Hardest Problems

On October 13, 2026, Koray Kavukcuoglu (SVP, Google DeepMind and Chief AI Architect, Google) officially introduced Gemini 4 Argon—Google's flagship frontier reasoning model. Designed specifically for deep work on complex, real-world tasks across coding, finance, legal, cybersecurity, and science, Argon is now globally available for developers, enterprises, scientists, and consumers.

Gemini 4 Argon is accessible via the API in Google AI Studio and the Gemini platform, for Google AI Ultra subscribers in Google Antigravity, and for Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search. The model launched at an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens discounted at 95%.

Its depth and precision across multi-step tasks leads Argon to outperform other frontier models on benchmarks across domains, ranking #1 on the Vals Index, which measures model performance across critical economic sectors like finance, coding, legal, and tax work.

At Google, teams have been deploying Argon on their hardest engineering problems. Over recent weeks, software engineers have been shipping high-quality code faster than ever. Inside Google Antigravity, Argon has helped teams increase agentic code submissions by more than 35% while reducing code rollbacks by nearly a third.

Untangling browser feature code in Chrome: Argon helped the Chrome team separate core webpage handling from a ~6 million line legacy folder of browser features held within Chromium’s 40 million-line codebase. Piping multiple clean code changes at a time, Argon helped the team finish an architectural refactor across hundreds of files, completing months of work in weeks. For Chrome's billions of global users, this unlocks the ability for live webpages to move fluidly between tabs, side panels, and floating windows without reloading, eliminating an entire category of browser crashes.

Optimizing a production Python codebase in hours: Argon in Antigravity is helping the team behind Google's open-source Agent Development Kit (ADK), a framework powering enterprise agents, push performance engineering further. The team gave Argon their .agents/skills rulebook and a one-paragraph goal to eliminate framework overhead without breaking existing apps. Working autonomously overnight, Argon surfaced optimizations, now live in google/adk-python, that cut in-memory session prep time by 40% in long sessions and sped up cold imports by 21%.

Quantum chemistry open-source package optimization: xTB (Semiempirical Extended Tight-Binding) is one of the most-cited open-source quantum chemistry packages for molecules. Scientists use it to predict molecules' 3D shapes and screen large numbers of drug and materials candidates. Using Argon in Antigravity, researchers made xTB up to 2.3x faster at predicting molecular structures and 1.8x faster at conformer search. The changes have been contributed to the official xTB repository on GitHub.

Making open Gemma models nearly three times as fast: Using Argon, Google achieved a 2.85x end-to-end inference speedup on an A100 via vLLM (accelerating throughput from ~37,900 to over 108,000 tokens per second on Gemma 4 E2B) with 100% token parity and zero quality drop in just a few days, a process that would usually take months of specialized kernel engineering.

Argon brings major improvements in spatial reasoning, mathematical precision, and visual coding compared to previous Gemini models. Using slash commands like /goal and prompting for multi-agent loops with custom skills in Antigravity, Googlers are using Argon to prototype complex 3D visualizations, high-fidelity games, and interactive animations in vanilla code that can be rendered directly in the browser.

Temple complex procedural modeling: Argon built a highly detailed model of a Pagoda using multi-agent visual critique loops in Antigravity. The model autonomously sculpted a multi-building temple complex with photorealistic architecture, foliage, and lighting in Blender.

Playable miniature racing game: Argon turned pictures of a real desk into a playable miniature racing game using a multi-agent loop in Antigravity. The desk objects, room, and race track were modeled and collision-checked using Blender MCP, textured with Nano Banana, and wired with physics, custom shaders, and car engine audio from Gemini Omni Flash in Three.js.

Procedural 3D game in WebGL: Argon in Antigravity wrote Python scripts to procedurally sculpt, rig, animate, and PBR-bake a complex 3D game using multi-agent visual critique loops, then wired it for the web in Three.js with GLSL shaders, shatter physics, and a single-player AI.

Pure Python 3D ray-tracing engine: With zero external libraries and 100% pure Python, Argon wrote its own 3D ray-tracing engine with a custom video tool and audio mixer to build and document a sunlit library from scratch, producing, narrating, and assembling a presentation video with Gemini 3.8 TTS voiceover and music by Lyria 3.5.

Gothic cathedral 3D synthesis: Starting from a single Nano Banana concept image in Antigravity, Argon drives Blender MCP in an autonomous loop, aligning camera grids, sculpting Gothic architecture, and staging audio-reactive lighting to create a finished 3D scene.

Argon combines frontier reasoning, deep visual perception, expanded domain knowledge, and stronger writing capabilities across legal, finance, science, and coding to help with high-stakes professional workflows. It communicates clearly, keeps the user in the loop, and follows style instructions better than any previous Gemini model. Argon answers questions directly with less hedging, doesn’t repeat itself unnecessarily, and skips clichés throughout responses.

Enterprise customers and startups in the early access program report significant operational wins across diverse sectors:

Using its advanced reasoning and multimodal capabilities while maintaining strict grounding, Argon excels in deep multimodal document comprehension, longitudinal data analysis, and interactive analytical sandboxes—such as synthesizing ~50 GB of LiDAR and CAD scans for archaeological reconstructions in Chaco Canyon, automated client proposal decks, and 3D printable mechanical prototyping.

Researchers, engineers, and security specialists are already using Argon to push the boundaries of cyber defense, science, math, and robotics. It’s helping cyber defenders and researchers test hypotheses, autonomously audit critical codebases and formal proofs, and catch subtle zero-day vulnerabilities and logic flaws.

State-of-the-art cyber capabilities with Wiz: Wiz is using Argon for its 'Scan for Good' initiative, protecting critical infrastructure for the public good. In collaboration with a federal agency, Argon proactively identified and remediated a gap that risked exposing sensitive government infrastructure. Wiz also used Argon to protect vital healthcare systems, chaining complex vulnerabilities to discover and patch a public file download bypass that exposed a master key and administrator access, safeguarding private patient records.

Quantum computing error correction: Argon translated an internal version of AlphaQubit, Google's ML quantum error correction decoder, into a single, highly optimized hardware kernel—nearly doubling its inference throughput to scale onto real quantum hardware in real time. For arcsin, a building block in quantum simulation algorithms, an implementation written by Argon and formally verified in Lean cuts spacetime cost by more than 100x compared to a published quantum CORDIC baseline.

Nuclear fusion and Grad's conjecture: Stellarators are leading candidates for clean fusion energy, but optimizing their 3D magnetic fields has been famously difficult. Since 1967, 'Grad’s conjecture' assumed exact mathematical solutions were impossible. Working alongside Professor Matt Landreman (University of Maryland), Google researchers used Argon to discover a broader class of counterexamples, providing the fusion community with reliable benchmarks for stellarator simulation codes.

Rare disease research at Gamow Labs: Researchers analyzed de-identified genomes from newborns with rare developmental lung disorders. Argon pinpointed a chromosomal rearrangement that routine software missed for years and explained why it went undetected, scoring 57% on RareBench and 52.2% on LabBench.

De novo enzyme design: Engineering serine esterases (enzymes that break down plastic ester bonds and assist pharmaceutical synthesis), Argon assembled relevant biochemical data, tuned candidate models, and applied biochemistry filters to produce five novel designs in hours. Subsequent laboratory testing confirmed robust catalytic activity across all five candidates.

Speeding up robotics development: On the Embodied Reasoning (ERQA) benchmark, Argon led frontier models with an 87% score. Testing out-of-distribution physical skill acquisition, Argon watched a raw video of a human playing cornhole, derived whole-body joint coordinates, built simulation code, and transferred the policy to Apptronik’s Apollo humanoid robot, which quickly sank seven consecutive shots in a real-world trial.

Safety and security are foundational to model development at Google. Because Argon is a highly capable model, Google collaborated with third-party partners to measure capabilities and enforce rigorous safeguards under the Frontier Safety Framework:

Misuse & jailbreak defenses: A multi-layer defense stack combines jailbreak resistance, monitoring, and enforcement. External bio red-teaming unlocked under 1% of tested bio sub-tasks, and violations on adversarial stress tests were reduced nearly 2-fold compared to previous Gemini models.

Agent alignment & real-time monitoring: To prevent autonomous agents from taking unauthorized shortcuts (such as ignoring failing tests or bypassing permission gates), Argon was trained to respect system boundaries, report negative findings honestly, and pause for human confirmation before taking irreversible steps. On coding benchmarks, shortcut attempts dropped to 2%. An automated misalignment monitor reads step-by-step reasoning and actions in real time without penalizing internal thoughts during training.

Prompt injection resistance: On the Gray Swan Indirect Prompt Injection (IPI) benchmark, Argon leads frontier models by reducing attack success across 15 adaptive attempts to 0.7%, hardening agents when reading untrusted web pages, emails, or codebases.

Gemini 4 Argon is rolling out globally with a 1,048,576 unified token context window via the Gemini API, featuring long-decode output generation and structured tool execution. Access is structured across developer, enterprise, consumer, and scientific tiers:

Developers can build with the Gemini API paid tier in AI Studio and Android Studio today. For deep dives into benchmark comparisons and coding workflows, explore our companion reviews:

• Gemini 4 Argon Review: Features, Pricing & Limitations — A breakdown of API pricing, coding performance, and early user feedback.

• Gemini 4 Argon vs Claude Opus 5.5: Coding Compared — Head-to-head comparison of context windows, pricing, agent harnesses, and benchmark rubrics.