Independent guide · Updated Sept 30, 2026

Gemini 4 Argon

Google just announced Gemini 4 Argon, its new frontier model for coding, knowledge work and cyber defense. Here is everything that is actually confirmed so far: the benchmarks, the price, the API and when you can use it.

Announced Sept 30, 2026
Public API: not yet
1M-token output

Gemini 4 Argon API See the benchmarks

Not affiliated with Google. Every number below comes from Google's announcement or named press, linked at the bottom.

Gemini 4 Argon at a glance

The short version, as of September 30, 2026.

AnnouncedSept 30, 2026On Google's blog, by Koray Kavukcuoglu of Google DeepMind

Built forCoding, knowledge work, cyber defenseLong, multi-step "agentic" jobs

Output limit1M tokensUp from 64K on earlier Gemini models

Price (intro)$2 in / $10 outPer million tokens. Later $4 / $20

API statusNot public yetNo model ID published so far

Who gets it firstCyber defendersThen AI Ultra and paid API customers, "soon"

What is Gemini 4 Argon?

Gemini 4 Argon is Google's newest frontier AI model, and the first model of the Gemini 4 generation. Google DeepMind announced it on September 30, 2026, and describes it as a model for real-world coding, enterprise knowledge work and cyber defense: the long, messy jobs that take many steps, many files and a lot of reading.

The name is new too. Instead of the familiar Pro and Flash labels, this one is called Argon, which press coverage describes as a new naming scheme for Gemini models. Google has not explained what "Argon" stands for, so we won't guess.

Three things stand out in the announcement:

A huge output limit. Argon can write up to 1 million tokens in a single answer, up from 64K. That is whole codebases, long reports or full migrations in one go.
It sees more than text. It takes text, images and video as input, including charts and visual documents, and Google reports a 91.7% score on a long-video understanding test.
A slow, careful rollout. Google says a model this capable needs a phased release, so trusted cyber defenders get it first and everyone else follows.

Want the details? Read when you can get it, what it costs, or how to use it.

Gemini 4 Argon benchmarks

Google's own numbers. No independent leaderboard has scored Argon yet.

All benchmarks →

DeepSWE v1.1

Real software engineering tasks. Higher is better.

Gemini 4 Argon77.9%

Claude Opus 5.574.2%

GPT-6 Astra74.1%

Claude Fable 5.167.4%

Across Google's full table of 19 tests, Argon is first on 13, ties on 1 and trails on 5. See where it loses.

LVBench91.7%Understanding long videos

CWE-bench v168%Finding security weaknesses, tied for first

AutomationBench51.3%Automating business workflows. Next best: 42.5%

Harvey's Legal Agent19.6%Legal work. Next best: 6.7%

For developers

Gemini 4 Argon API

This is the question everyone is asking, so here is the straight answer: the Gemini 4 Argon API is not open to the public yet. Google says it is coming "soon" to paid API customers, right after trusted security partners in its Fairwind Program. There is no public model ID yet, and Argon is not on the Gemini API model list today.

Input$2 / 1MIntro price. Standard: $4

Output$10 / 1MIntro price. Standard: $20

Cached input95% offAbout $0.10 per 1M at the intro price

You can be ready on day one. Get a Gemini API key in Google AI Studio, build against today's model, and keep the model name in one setting so switching to Argon is a one-line change:

# pip install google-genai
import os
from google import genai

client = genai.Client()  # reads GEMINI_API_KEY from your environment
MODEL = os.environ.get("GEMINI_MODEL", "gemini-3.8-flash")  # swap in the Argon ID when Google publishes it

response = client.models.generate_content(model=MODEL, contents="Plan a migration of this repo to Rust.")
print(response.text)

Read the full Gemini 4 Argon API guide for pricing math, the OpenAI-compatible endpoint, JavaScript and curl examples, and what to watch for at launch.

While you wait for Argon

Make videos and images with Google's best models today

Argon is a thinking and coding model: it reads video, it doesn't make it. Google's video makers are Veo 3.1 and Gemini Omni, and its top image model is Nano Banana 2. You can use all three right now on Lumeta, next to Seedance, Kling and more, in one account.

Every top image, video, voice and music model in one place
One account and one set of credits, instead of five subscriptions
Made to cost less than Higgsfield, Magnific and Artlist

Try Lumeta

Full disclosure: Lumeta is made by the team behind this site.

Veo 3.1Google's flagship video model, with sound Gemini OmniCreate and edit videos by chatting Nano Banana 2Google's image maker and editor Seedance 2.5Cinematic video from text or images Kling O3Video with control over every character Upscale VideoSharpen any clip up to 4K

Gemini 4 Argon guides

Short, sourced answers to the questions people are searching right now.

DevelopersGemini 4 Argon APIModel ID status, pricing and code you can run today. ScoresBenchmarks vs Claude and GPTHow Argon stacks up against Claude Opus 5.5 and GPT-6 Astra. CostPricing and cost calculator$2 / $10 per million tokens, and what a real job costs. AccessRelease date and who gets itFairwind, AI Ultra, the API, and whether it's free. How toHow to use Gemini 4 ArgonEvery route in, step by step, plus prompting tips. How toConnect Gemini 4 Argon to ClaudeUse Gemini inside Claude Code and Claude Desktop with MCP. How toMaking videos with Gemini 4 ArgonWhat Argon can and can't do with video, and what to use instead. TodayAlternatives you can use nowThe best options while Argon rolls out. AnswersGemini 4 Argon FAQEvery common question, answered in a sentence or two.

Gemini 4 Argon FAQ

What is Gemini 4 Argon?

Gemini 4 Argon is Google's newest frontier AI model, announced on September 30, 2026. Google built it for real-world coding, enterprise knowledge work and cyber defense, and it can write up to 1 million tokens in one answer.

When is the Gemini 4 Argon release date?

It was announced on September 30, 2026, but it is rolling out in phases. Trusted cyber defenders get it first through Google's Fairwind Program. Google AI Ultra subscribers and paid API customers are next, "soon", with no exact date yet.

Is the Gemini 4 Argon API available?

Not publicly yet. Google has published the price ($2 per million input tokens and $10 per million output tokens at launch) but not a model ID. Paid API customers are among the next groups to get access.

Is Gemini 4 Argon free?

Google has not announced a free tier. The first consumer access is planned for Google AI Ultra, Google's top paid plan.

Can Gemini 4 Argon make videos?

Google has not said it generates video. Argon understands video very well, but Google's video makers are Veo 3.1 and Gemini Omni, which you can use today on platforms like Lumeta.

See all Gemini 4 Argon questions →

Sources: Google's announcement, 9to5Google, The New Stack, Gemini API models.