Google announced the global release of Gemini 4 Argon on October 13, 2026, launching its latest frontier artificial intelligence model across enterprise platforms, developer tools, and consumer applications.
Availability and Pricing
According to Google, Gemini 4 Argon is available globally for developers and enterprises through the API in AI Studio and the Gemini platform, as well as for Google AI Ultra and Pro subscribers in the Gemini app, AI Mode in Search, and Google Antigravity.
The company stated that the model carries an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens discounted by 95%. Google added that the model supports a 1 million unified input-output token limit via the Gemini API.
Performance and Benchmark Results
Google claimed Gemini 4 Argon ranks first on the Vals Index, which measures model performance across sectors including finance, coding, legal, and tax work. On the Embodied Reasoning benchmark evaluating physical and spatial reasoning in 3D environments, Google reported the model achieved an 87% score.
In evaluations conducted by Gamow Labs on real clinical and laboratory data, the model scored 57% on RareBench for rare-disease diagnosis assistance and 52.2% on LabBench for wet-lab reasoning, according to the announcement. Google also reported that shortcut attempts on coding benchmarks dropped to 2%.
Engineering and Security Applications
Google reported that its internal software engineering teams used the model in Google Antigravity to increase agentic code submissions by more than 35% while reducing code rollbacks by nearly a third.
In cybersecurity, Google announced the expansion of access to hundreds of partners in its Fairwind Program. Early access partner Wiz used the model to identify infrastructure vulnerabilities, including uncovering a public file download bypass that exposed administrator access and a master key at a healthcare organization, Google stated.
On safety testing, Google said the model reduced attack success on the Gray Swan Indirect Prompt Injection benchmark to 0.7% across 15 adaptive attempts.
What to do
- Review Google developer and enterprise documentation on the new continuation protocol to ensure proper configuration when implementing the Gemini API.
- Eligible researchers can apply for six months of free access to the highest usage tier in Google Antigravity.
- Cybersecurity defenders can apply to Google's Fairwind Program to request access to Argon without standard cyber guardrails.
Key facts and where they come from
- Gemini 4 Argon launched at an introductory price of $2 per million input tokens and $10 per million output tokens.
The model launches today at an introductory price 1 of $2 per million input tokens and $10 per million output tokens
- Google reported that Gemini 4 Argon achieved an 87% score on the Embodied Reasoning benchmark.
Relying on pure reasoning, Argon outperforms all other frontier models, with an 87% score.
- The model supports a 1 million unified input-output token limit through the Gemini API.
This is enabled by an industry-leading 1M unified input-output token limit via the Gemini API.
- Argon reduced attack success on the Gray Swan Indirect Prompt Injection benchmark to 0.7% across 15 adaptive attempts.
Argon leads the Gray Swan Indirect Prompt Injection (IPI) benchmark, cutting attack success across 15 adaptive attempts to 0.7%.
