Is Gemini 4 Argon out?

Gemini 4 Argon is announced but not generally available. Google is rolling Argon out first to trusted cyber defenders through its Fairwind Program. Developers, enterprises and consumers come later, and Google has not given a date.

When will Gemini 4 Argon be available to developers?

Google says it will make Gemini 4 Argon available to developers, enterprises and consumers as soon as possible after early testing, starting with paid API customers and Google AI Ultra subscribers. Google has not announced a date. reAPI will update this page if Argon becomes available here.

Is Gemini 4 Argon free or paid?

Paid. Google has announced introductory per-token pricing for Gemini 4 Argon, with a higher standard rate after the introductory period, and cached input tokens priced well below regular input. There is no free tier announced. Argon is not on reAPI yet, so this page shows no reAPI rate.

Does Gemini 4 Argon require a monthly subscription?

Not for API use. Google prices Gemini 4 Argon per token for API customers. Consumer access is set to start with Google AI Ultra subscribers, which is a Google subscription. Chat models on reAPI are pay-as-you-go per token with no monthly plan.

How do I get a Gemini 4 Argon API key?

No one can get a Gemini 4 Argon API key yet, because Google has not opened developer access. You can create a reAPI API key in your dashboard today and use it with the chat models already live on reAPI, such as Claude Opus 5.5 and Gemini 3.6 Flash.

What is the Gemini 4 Argon output limit?

Gemini 4 Argon has a 1M-token output limit, up from 64K. Google says that headroom lets the model think deeply and produce hundreds of thousands of tokens in one trajectory. Google has not yet published the Argon context window.

How does Gemini 4 Argon perform on benchmarks?

In Google's table, Gemini 4 Argon scores 77.9% on DeepSWE v1.1, 51.3% on AutomationBench, 91.7% on LVBench and 68% on CWE-bench v1, and leads the Vals Index. GPT-6 Astra or Claude Opus 5.5 lead on some tests, including FrontierSWE v2, Terminal-bench 4.0 and OSWorld-2.0.

Is Gemini 4 Argon better than Claude Opus 5.5?

It depends on the task. Gemini 4 Argon beats Claude Opus 5.5 on DeepSWE v1.1, legal work and long video in Google's table, while Claude Opus 5.5 leads on Terminal-bench 4.0 and PostTrainBench. Claude Opus 5.5 is available on reAPI now; Argon is not.

Can I run Gemini 4 Argon locally?

No. Google has announced Gemini 4 Argon as a hosted model delivered through its own programs and API, and has not announced downloadable weights. There is no official Argon download or open-source release.

What is the Fairwind Program?

Fairwind is Google's program that gives trusted cyber defenders early access to frontier models. Google prioritizes governments, critical infrastructure operators and core technology platforms, vets every applicant, and works with over 650 partners. A set of them get Gemini 4 Argon access, without cyber guardrails, for vulnerability research and patching.

Gemini 4 Argon — Google's Frontier Model for Complex Work

Gemini 4 Argon is Google's new frontier model for long-horizon coding, legal and finance research, and cyber defense, with a 1M-token output limit. Google is releasing Argon in phases, and it is not on reAPI yet.

Coming soon — the API for this model is not available yet.

Highlights: 1M-token output · long-horizon reasoning · limited preview
Best for: Codebase migrations, legal and finance research, security reviews
Input: Text, documents, charts and long video, per Google
Output: Text, up to 1M tokens per response

What is Gemini 4 Argon?

Gemini 4 Argon is the frontier model Google announced on September 30, 2026. Google built Argon to sustain deep reasoning across long, multi-step workflows: real-world software engineering, enterprise knowledge work such as legal and finance, and cybersecurity defense. The headline change is output length. Argon can return up to 1M tokens in one response, up from 64K, so a single run can carry a large migration or a long report to the end. Access is phased: Argon is rolling out first to trusted cyber defenders in Google's Fairwind Program, and Google says developers, enterprises and consumers come next, starting with paid API customers. Argon is not available on reAPI yet; this page tracks what Google has confirmed.

Gemini 4 Argon poster: long-horizon coding and research on reAPI

What you can build with this model

Real-world workflows and production use cases you can build and ship with this model.

Gemini 4 Argon poster for long codebase migrations

Carry a large codebase migration to the end

Google engineers already use Gemini 4 Argon for debugging and large-scale codebase migrations. Google reports Argon agents moving C and C++ code to Rust across its own projects, up to 800K+ lines for the Fuchsia Zircon kernel, with automated and manual review before anything ships. In libgav1, Argon replaced 32K lines of SIMD code with safe Rust that runs 2.7x faster than the earlier Rust port. The 1M-token output limit is what lets Argon keep going on work of this size.

Gemini 4 Argon poster for legal and finance research

Draft legal and finance research

Gemini 4 Argon leads the Vals Index in Google's announcement, a benchmark that weights finance, coding, legal and tax work by its share of U.S. GDP. Google also reports Argon at the top of Vals Finance Agent v2 for multi-step financial research and Harvey's Legal Agent Benchmark for legal research and drafting. Argon can read charts, long videos and a series of documents, which matters when the evidence sits in filings, decks and recordings rather than plain text.

Gemini 4 Argon poster for defensive security reviews

Find and patch vulnerabilities

Google trained Gemini 4 Argon for defensive security work: Google says Argon can autonomously find, validate and patch critical software vulnerabilities. Argon ties for first on CWE-bench v1, which measures vulnerability remediation, and Wiz is already using it in its Scan for Good program to protect public infrastructure. This is why Argon reaches cyber defenders first.

Why reAPI for Gemini 4 Argon

Get your reAPI key ready

Google has not opened Gemini 4 Argon to developers yet. A reAPI key already works for the frontier chat models live on reAPI today, including Claude Opus 5.5 and Gemini 3.6 Flash, so you can build and test your workflow now.

One chat endpoint across models

reAPI's live chat models share one chat completions endpoint and request format. Switching models in an existing workflow is a change to the model field, which keeps a later Gemini 4 Argon evaluation small.

Pay per token, no subscription

reAPI chat models are billed per token, pay-as-you-go, with no monthly plan. Live per-token rates for each model are on its model page; Gemini 4 Argon has no reAPI rate until it launches here.

Gemini 4 Argon vs Claude Opus 5.5

Claude Opus 5.5 is one of the three models Google compares Gemini 4 Argon against. Argon leads on long-horizon coding, legal work and long video in Google's table; Claude Opus 5.5 leads on terminal work and is available on reAPI today.

Capability

Gemini 4 Argon

Claude Opus 5.5

Availability

Limited rollout to Fairwind Program cyber defenders; not on reAPI yet

Available on reAPI now

Maximum output

1M tokens

128K tokens

Context window

Not yet published by Google

1M tokens

DeepSWE v1.1 (long-horizon coding)

77.9%

74.2%

Terminal-bench 4.0 (terminal work)

57.4%

66.4%

Harvey's Legal Agent Benchmark

19.6%

3.8%

LVBench (long video)

91.7%

83.7%

Scores come from the benchmark table in Google's Gemini 4 Argon announcement (September 30, 2026); per Google's methodology, non-Gemini scores are mostly self-reported by each provider. Claude Opus 5.5 limits are from its reAPI model page. Checked October 1, 2026.

How to prepare for Gemini 4 Argon

step 01
Create a reAPI API key
Sign up and create an API key in your dashboard. The same key covers every chat model live on reAPI.
Open
step 02
Run your workload on Claude Opus 5.5
Claude Opus 5.5 is a frontier model in Google's own Gemini 4 Argon comparison and is live on reAPI today. Use it to build and benchmark your workflow now.
Open
step 03
Learn the chat request format
Read the chat completions docs. When Gemini 4 Argon reaches reAPI, the evaluation should be a change to the model field.
Open

docs/api/gemini-4-argon

API reference

Drop-in code and the full parameter table.

View full reference

curl https://reapi.ai/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-4-argon",
    "messages": [
      { "role": "user", "content": "Plan a C++ to Rust migration for this module" }
    ]
  }'

Parameter validation failed

Rejected synchronously with the failing field named. Check enum values, ranges, and URL fields — nothing is charged.

Authentication or balance issue

401 means a missing or invalid Bearer key; 402 means the reserve exceeds your balance. Manage keys and credits in the console.

Content or material rejected

Reference material and outputs pass automated moderation. Rejected tasks fail with a clear error and a full refund.

Task failed or timed out

A task that reaches a failed state is never charged — the reserve refunds automatically. Keep the task id and retry when ready.

Frequently asked questions

Common questions about this model.

Gemini 4 Argon is Google's frontier model, announced on September 30, 2026. Argon targets long-horizon software engineering, legal and finance knowledge work, and cybersecurity defense, and it can output up to 1M tokens in a single response.