Gemini 4 Argon: Google DeepMind's Frontier Model for Coding, Science, and Agentic Work

Google DeepMind announced Gemini 4 Argon on October 13, 2026, a frontier model positioned for deep, multi-step work in coding, finance, legal, cybersecurity, and science. The release matters because Argon is not pitched as a general chat upgrade but as a model for long-horizon agentic tasks, and DeepMind backs that claim with internal engineering results, third-party security and science deployments, and a published safety stack.

Availability and pricing

Gemini 4 Argon is globally available as of October 13, 2026, across developer, enterprise, and consumer surfaces. DeepMind describes it as the company's latest frontier model, released to developers, enterprises, scientists, and consumers simultaneously.

Developers: Available via the API in AI Studio and the Gemini platform, and in the Gemini API paid tier in AI Studio and Android Studio. Google AI Ultra subscribers can also try it in Google Antigravity and AI Studio.
Enterprises: Available in Gemini for business, the API, and Google Antigravity.
Consumers: Available in the Gemini app and AI Mode in Search for Google AI Ultra and Pro subscribers.
Scientists and researchers: DeepMind is offering 10,000 scientists the highest usage tier in Google Antigravity free of charge for 6 months, onboarded on a rolling basis and subject to eligibility, capacity, and selection.
Cybersecurity defenders: Applications are open to join the Fairwind Program to get access to Argon without cyber guardrails on a rolling basis.

Argon launches at an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95% off. The model supports an industry-leading 1M unified input-output token limit via the Gemini API, enabled by a new continuation protocol documented for developers and enterprises.

Benchmark positioning

DeepMind states that Argon's depth and precision across multi-step tasks lead it to outperform other frontier models on benchmarks across domains. The company cites a #1 ranking on the Vals Index, which measures model performance across critical economic sectors including finance, coding, legal, and tax work.

On the Embodied Reasoning (ERQA) benchmark, which tests a model's ability to understand and solve physical, spatial problems in 3D environments, Argon scores 87% and outperforms all other frontier models relying on pure reasoning.

Internal engineering results at Google

DeepMind reports that Google software engineers have been using Argon on the company's hardest engineering problems, and that Argon in Antigravity helped teams increase agentic code submissions by more than 35% while reducing code rollbacks by nearly a third.

Chrome: refactoring a 40-million-line codebase

Argon helped the Chrome team separate core webpage handling from a roughly 6-million-line legacy folder of browser features inside Chromium's 40-million-line codebase. Piping multiple clean code changes at a time, Argon helped the team complete an architectural refactor across hundreds of files, finishing months of work in weeks. For Chrome's billions of global users, DeepMind says this will unlock the ability for live webpages to move fluidly between tabs, side panels, and floating windows without reloading, eliminating an entire category of browser crashes.

Agent Development Kit: overnight Python optimization

The team behind Google's open-source Agent Development Kit (ADK), a framework powering enterprise agents, 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%.

xTB: quantum chemistry speedups

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

Gemma and vLLM: 2.85x inference speedup

Using Argon, DeepMind achieved a 2.85x end-to-end inference speedup on an A100 via vLLM, accelerating throughput from roughly 37,900 to over 108,000 tokens per second on Gemma 4 E2B, with 100% token parity and zero quality drop. The work took a few days rather than the months it would usually require, and the changes were submitted as upstream pull requests covering attention tuning and speculative decoding.

3D visualization and visual coding

DeepMind reports major improvements in spatial reasoning, mathematical precision, and visual coding compared with previous Gemini models. Using slash commands such as /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 renders directly in the browser.

Professional and enterprise workflows

Argon combines frontier reasoning, deep visual perception, expanded domain knowledge, and stronger writing capabilities across legal, finance, science, and coding to support high-stakes professional workflows. DeepMind says the model communicates clearly, keeps the user in the loop, and follows style instructions better than any previous Gemini model. It answers questions more directly with less hedging, does not repeat itself more than necessary, and skips clichés.

Enterprise customers and startups in the early access program are already reporting wins, according to DeepMind. The company says Argon excels in deep multimodal document comprehension, longitudinal data analysis, and interactive analytical sandboxes while maintaining strict grounding.

Cybersecurity: Fairwind expansion and Wiz deployments

DeepMind is beginning expansion of the Fairwind Program to hundreds of additional partners. Wiz is already using Argon for its "Scan for Good" initiative, which protects critical infrastructure for the public good and for free. With early access to the model, Wiz demonstrated two high-stakes results:

Securing sensitive government data: As part of a long-standing cyber resilience program and in collaboration with a federal agency, Argon helped proactively identify and remediate a gap that risked exposing sensitive data and government infrastructure. DeepMind says the effort helped prevent exposure of critical security sector development details and unauthorized access to specialized models.
Protecting critical healthcare infrastructure: Wiz used Argon to protect vital healthcare infrastructure. By chaining complex vulnerabilities, the model helped the team discover and quickly fix a public file download bypass that exposed a master key and granted administrator access. Argon went beyond spotting a vulnerability to proactively safeguarding hundreds of thousands of private patient records.

Scientific breakthroughs

DeepMind lists several domains where Argon is already producing results for researchers and domain experts.

Quantum computing

Argon is helping DeepMind researchers accelerate progress in quantum error correction and algorithm design. To help error decoders keep pace with quantum hardware in real time, Argon translated an internal version of AlphaQubit, DeepMind's ML quantum error correction decoder, into a single highly optimized hardware kernel, nearly doubling its inference throughput so the team can deploy larger, more capable models and scale to larger quantum devices while meeting strict real-time thresholds. Argon is also reducing spacetime costs for quantum subroutines: for arcsin, a building block in recent quantum simulation algorithms, an implementation written by Argon and formally verified in Lean cuts spacetime cost by more than 100x compared with a published quantum CORDIC baseline.

Nuclear fusion

Stellarators are complex machines that represent a leading candidate for limitless clean fusion energy, but optimizing their 3D magnetic fields is famously difficult. Since 1967, a long-standing hypothesis known as "Grad's conjecture" assumed that finding exact mathematical solutions for these fields was impossible, until recent counterexamples were found, including by Professor Matt Landreman of the University of Maryland. A team at Google alongside Professor Landreman has now discovered a broader class of counterexamples to Grad's conjecture. These new solutions give the fusion community reliable benchmarks to test stellarator simulation codes.

Rare disease research

Researchers at Gamow Labs used Argon to analyze de-identified genomes from newborns with rare developmental lung disorders. In one case, Argon pinpointed a chromosomal rearrangement that routine software missed for years and explained why it went undetected. On private Gamow Labs evaluations derived from real clinical and lab data, Argon topped RareBench with 57% on assistance with rare-disease diagnosis and LabBench with 52.2% on wet-lab reasoning.

Enzyme design

Engineering enzymes with precise reactivity is notoriously slow, often demanding months of bespoke model tuning and candidate screening. Serine esterases are a class of enzymes that cut ester bonds on targeted molecules, widely used in recycling and in the manufacture of pharmaceuticals, biofuels, and detergents. DeepMind scientists worked with Argon to design serine esterases targeting the benchmark molecule 4MU-acetate, which glows under UV light once its ester bonds are broken. In a matter of hours, Argon assembled relevant data, performed model tuning, and identified multiple biochemistry-inspired filters to reach a final set of five novel designs. Subsequent laboratory testing confirmed robust catalytic activity across these designs.

Robotics: video-to-sim-to-real

DeepMind says Argon brings together advanced visual-semantic reasoning, deep physical understanding, and autonomous code generation to unlock new workflows across the robotics stack. In an internal test, Argon watched a raw video of a human playing cornhole, mapped the spatial coordinates of the environment, and calculated the required whole-body joint movements. It then autonomously built the simulation code. When transferred onto Apptronik's Apollo humanoid, the robot sank seven consecutive shots in a real-world trial. DeepMind calls this pipeline "Video-to-Sim-to-Real" and says it compresses weeks of manual simulation modeling into a few hours.

Safety and alignment

DeepMind describes Gemini 4 Argon as its most comprehensive safety and security stack yet, developed in collaboration with third-party partners on measurement of certain model capabilities and safeguards.

Misuse and jailbreak defenses: Under the Frontier Safety Framework, DeepMind evaluated Argon's enhanced capabilities across several domains and deployed a multi-layer defense stack combining jailbreak resistance, monitoring, and enforcement. External bio red teaming could only unlock less than 1% of the bio sub-tasks tested for. On an adversarial set meant to stress-test the model, violations reduced nearly 2-fold, which DeepMind says is significantly better than previous Gemini models.
Agent alignment and real-time monitoring: DeepMind trained Argon to respect system boundaries, report negative findings honestly, and pause for human confirmation before taking irreversible steps. On coding benchmarks where models have no incentive to hold back, shortcut attempts dropped to just 2% on Gemini 4 Argon. As a defense-in-depth layer, DeepMind built an automated misalignment monitor that reads the agent's step-by-step reasoning and actions to catch and block unintended workarounds in real time, used across internal deployments, capability evaluations, and reinforcement learning training runs, while avoiding penalizing the model's thoughts during training so its reasoning remains transparent and monitorable.
Prompt injection resistance: Through adversarial training and automated red teaming, Argon leads the Gray Swan Indirect Prompt Injection (IPI) benchmark, cutting attack success across 15 adaptive attempts to 0.7%.

A full Gemini 4 Argon Model Card is published alongside the announcement.

What comes next

DeepMind frames Gemini 4 Argon as the beginning of the Gemini 4 family, with the rest of the lineup to follow. The model's combination of a 1M unified input-output token limit, a new continuation protocol for long decodes, and reported gains in agentic reliability positions it as a direct competitor for long-horizon coding, scientific, and enterprise agent workloads.