Google DeepMind · 2026-09-30 · seismic
Gemini 4 Argon — Google's new frontier model gets a 1M-token output limit
Gemini 4 Argon is Google's first Gemini 4 model. It scores 77.9% on DeepSWE v1.1, ahead of Claude Opus 5.5 and GPT-6 Astra, and can write up to 1M output tokens. Cyber defenders get it first; the API follows.
Google's answer to Opus 5.5 and GPT-6 Astra, built for long coding, legal and finance jobs and for cyber defence.
Quick facts
| Maker | Google DeepMind |
|---|---|
| Max output | 1M tokens (was 64K) |
| Price (intro) | $2 in / $10 out per 1M tokens |
| Price (later) | $4 in / $20 out per 1M tokens |
| Availability | Fairwind Program first; paid API and AI Ultra next |
| Modalities | Text, code, images and charts, long video |
Benchmarks
Pricing
| Input (intro) | $2.00 / 1M tokens |
|---|---|
| Output (intro) | $10.00 / 1M tokens |
| Cached input · off the input price | 95% off |
| Input (standard) · after the intro period | $4.00 / 1M tokens |
| Output (standard) · after the intro period | $20.00 / 1M tokens |
What is it?
Gemini 4 Argon is Google's new frontier model and the first of the Gemini 4 generation, announced on September 30, 2026 by Koray Kavukcuoglu of Google DeepMind. Google built it for long, multi-step work: real-world software engineering, legal and finance research, and cybersecurity defence. Google says its engineers already use Argon every day for debugging, codebase migrations and algorithm design.
How does it work?
The biggest technical change is output length: the output token limit rises to 1M tokens, up from 64K, so one response can hold a very large piece of work. Release is phased. Trusted defenders in Google DeepMind's Fairwind Program get Argon first, without cyber guardrails, while the public version refuses cyber and CBRN attack requests. Google adds activation monitoring for misuse and chain-of-thought monitoring for misalignment.
Why does it matter?
Google's charts put Argon ahead of both rival flagships on long coding tasks and workflow automation, so the frontier race is three-way again after weeks of Anthropic and OpenAI launches. The 1M-token output limit matters for agents that migrate or rewrite whole codebases; Google reports internal C/C++-to-Rust migrations of up to 800K+ lines. Most developers must wait, though: there is no public API date yet.
Who is it for?
developers building long-running coding and knowledge-work agents, and security teams
Frequently asked questions
- When can developers use Gemini 4 Argon through the API?
- Google has not given a date. Gemini 4 Argon rolls out first to trusted cyber defenders in the Fairwind Program, and Google says it will reach paid API customers and Google AI Ultra subscribers as soon as possible. Google also says it is working with the U.S. government's voluntary pre-release process.
- How much will Gemini 4 Argon cost?
- Gemini 4 Argon has an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input 95% off the input price. After the introductory period, Google says the price becomes $4 per million input tokens and $20 per million output tokens.
- How does Gemini 4 Argon compare with Claude Opus 5.5 and GPT-6 Astra?
- In Google's own launch charts, Gemini 4 Argon scores 77.9% on DeepSWE v1.1 against 74.2% for Claude Opus 5.5 and 74.1% for GPT-6 Astra. On Zapier's AutomationBench it scores 51.3% against 42.5% and 41.4%. These are Google's numbers; independent tests are not out yet.
- What is the Fairwind Program?
- The Fairwind Program is Google DeepMind's vetted-access program for cyber defenders. It prioritizes governments, critical infrastructure operators in healthcare, telecom, energy and finance, and core technology platforms. Members must use Gemini 4 Argon only in internal security teams, use phishing-resistant MFA, and cannot share or resell access.
Try it
Apply through the Fairwind Program at https://deepmind.google/fairwind-program/ (API access for paid customers comes later)