By No Code MBA · Facts checked
- Provider
- Announcement date
- Status at last check
- Generally available
- Model family
- Gemini 4
- Capabilities
- Text, Image, Coding
What changed
Gemini 4 Argon is the first model Google has released in the Gemini 4 family, aimed at long-horizon work in software engineering, finance, legal, cybersecurity and science. Through the Gemini API it has a 1M unified input-output token limit. Google reports it ranks first on the Vals Index, scores 87% on the ERQA embodied reasoning benchmark, and has a 0.7% attack success rate on Gray Swan's indirect prompt injection benchmark. Google also says it follows style instructions better and hedges less than earlier Gemini models.
Access and availability
Use model ID gemini-4-argon on the paid Gemini API through Google AI Studio and the Gemini platform on Google Cloud, or in Android Studio on paid tiers. AI Ultra subscribers get it in Google Antigravity, AI Pro and Ultra subscribers in the Gemini app and AI Mode in Search, and vetted defenders can apply to the Fairwind Program for access without cyber guardrails.
Official access documentationPricing
$2 per million input tokens and $10 per million output tokens (introductory, cached input 95% off), rising to $4 input and $20 output per million tokens from November 30, 2026
Official pricingWhat you can build
If you have access, compare Argon with your current model on a representative long-horizon coding or document-work task, tracking completion quality and review effort rather than relying on provider benchmark claims.
Limitations
The current price is introductory and doubles on November 30, 2026. Consumer and IDE access is tied to paid subscription tiers, and many reported results come from Google, partners or private evaluations rather than public benchmarks.