Google has released Gemini 4 Argon, its newest model, which the company positions for deep, multi-step work in coding, finance, legal, cybersecurity, and science. Google says Argon outperforms other frontier models across several benchmark domains and ranks #1 on the Vals Index, which measures model performance across critical economic sectors such as finance, coding, legal, and tax work. The model is available today to developers and enterprises through the API in AI Studio and the Gemini platform, with access also extended to subscribers of several Google products.

Pricing and Access

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. Google AI Ultra subscribers can use it in Google Antigravity, and Google AI Ultra and Pro subscribers can use it in the Gemini app and AI Mode in Search.

Real Results Across Engineering

Google says teams inside the company and across enterprises and the research community have already used Argon for large codebase refactoring and optimization, business workflows in legal, finance, and science, finding and patching critical cybersecurity vulnerabilities, and early work in quantum computing, nuclear fusion, and life sciences.

On its own engineering work, Google says Argon in Antigravity helped teams increase agentic code submissions by more than 35% while reducing code rollbacks by nearly a third. Several examples stand out:

• Chrome refactor: Argon helped the Chrome team separate core webpage handling from a roughly 6 million line legacy folder of browser features held within Chromium's 40 million-line codebase. The model finished an architectural refactor across hundreds of files, completing months of work in weeks. Google says this will let live webpages move between tabs, side panels, and floating windows without reloading, eliminating an entire category of browser crashes.
• Agent Development Kit: The team behind Google's open-source Agent Development Kit (ADK) gave Argon their .agents/skills rulebook and a one-paragraph goal: 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 package: Google researchers used Argon in Antigravity to make xTB, one of the most-cited open-source quantum chemistry packages for molecules, 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.
• Gemma models: Using Argon, Google achieved a 2.85x end-to-end inference speedup on an A100 via vLLM, with throughput rising from about 37,900 to over 108,000 tokens per second on Gemma 4 E2B. Google reports 100% token parity and zero quality drop, achieved in just a few days, a process that would usually take months. The upstream pull requests cover optimizations from attention tuning to speculative decoding.
• 3D visualizations: Google says Argon brings major improvements in spatial reasoning, mathematical precision, and visual coding compared with previous Gemini models. Googlers are using it with slash commands like /goal and multi-agent loops with custom skills to prototype 3D visualizations, high-fidelity games, and interactive animations in vanilla code that renders directly in the browser.

Better Writing and Professional Workflows

Google says Argon combines frontier reasoning, deep visual perception, expanded domain knowledge, and stronger writing across legal, finance, science, and coding. The company says it communicates clearly, keeps users in the loop, and follows style instructions better than any previous Gemini model. Argon answers questions more directly with less hedging, avoids unnecessary repetition, and skips clichés. Enterprise customers and startups in the early access program are seeing strong results, according to Google, particularly in deep multimodal document comprehension, longitudinal data analysis, and interactive analytical sandboxes, all while maintaining strict grounding.

Cybersecurity Defense

Google says researchers and security specialists are using Argon to test hypotheses, autonomously audit critical codebases and formal proofs, and catch subtle zero-day vulnerabilities and logic flaws. Google is expanding access to hundreds of additional Fairwind Program partners.

Wiz is already using Argon for its “Scan for Good” initiative, which protects critical infrastructure for the public good and for free. Wiz reported two examples from its early access:

• Government data: In a long-standing cyber resilience program with a federal agency, Argon helped identify and remediate a gap that risked exposing sensitive data and government infrastructure, helping prevent exposure of critical security sector development details and unauthorized access to specialized models.
• Healthcare infrastructure: By chaining complex vulnerabilities, Argon helped the team discover and quickly fix a public file download bypass that exposed a master key and granted administrator access. Google says the model went beyond spotting the flaw to proactively safeguard hundreds of thousands of private patient records.

Scientific Breakthroughs

Google highlights several research results:

• Quantum computing: To help error decoders keep pace with quantum hardware in real time, Argon translated an internal version of AlphaQubit, Google's ML quantum error correction decoder, into a single, highly optimized hardware kernel. This nearly doubled its inference throughput, allowing larger, more capable models and scaling to larger quantum devices while still meeting strict real-time thresholds. 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: Optimizing the 3D magnetic fields of stellarators has been difficult. Since 1967, “Grad's conjecture” assumed that exact mathematical solutions for these fields were impossible, until recent counterexamples were found, including by Professor Matt Landreman of the University of Maryland. A Google team working with Landreman has found a broader class of counterexamples. Google says these solutions give the fusion community reliable benchmarks for testing 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 had missed for years and explained why. On private Gamow Labs evaluations built from real clinical and lab data, Argon scored 57% on RareBench for assistance with rare-disease diagnosis and 52.2% on LabBench for wet-lab reasoning.
• Enzyme design: Serine esterases cut ester bonds and are used in recycling, pharmaceutical manufacturing, biofuels, and detergents. Google 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 used biochemistry-inspired filters to reach a final set of five novel designs. Subsequent lab testing confirmed robust catalytic activity across these designs.

Google also plans to offer 10,000 scientists the highest usage tier in Google Antigravity free of charge for six months. Onboarding happens on a rolling basis and is subject to eligibility, capacity, and selection.

Key Features at a Glance

• Benchmark leadership: #1 on the Vals Index – Google's measure of performance across finance, coding, legal, and tax work, useful for judging fit for economically critical tasks.
• Introductory pricing: $2 per million input tokens, $10 per million output tokens – a clear cost basis for API planning.
• Cached input savings: Cached input tokens at 95% off – lowers cost for repeated context.
• Engineering throughput: More than 35% increase in agentic code submissions, with code rollbacks reduced by nearly a third – measurable gains in developer productivity.
• Inference speedup: 2.85x end-to-end on an A100 via vLLM, from about 37,900 to over 108,000 tokens per second on Gemma 4 E2B – with 100% token parity, so output is unchanged.
• Quantum error correction: AlphaQubit decoder throughput nearly doubled – supports larger models for quantum devices under real-time constraints.
• Rare disease evaluations: 57% on RareBench and 52.2% on LabBench in Gamow Labs' private tests – indicates performance on specialized clinical and lab reasoning, though these are vendor-run private evaluations.

Availability

Argon is available today through the API in AI Studio and the Gemini platform for developers and enterprises. Google AI Ultra subscribers can use it in Google Antigravity, and Google AI Ultra and Pro subscribers can access it in the Gemini app and AI Mode in Search. The introductory API price is $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95% off.