The headline says everyone. The body names paid tiers
Google's post today says Gemini 4 Argon is open to developers and enterprises through the API in AI Studio and the Gemini platform, to Google AI Ultra subscribers in Google Antigravity, and to Ultra and Pro subscribers in the Gemini app and AI Mode in Search. The developer section specifies the paid API tier. The post does not say whether a free account can use this model.
The shared input and output context limit is 1 million tokens. Google is also expanding the Fairwind program to hundreds of additional partners and taking applications for cyber defenders to use Argon without cyber guardrails. That is an access arrangement, not a published list of vulnerabilities.
[1]
The introductory price lasts through November 30
The launch price is $2 per million input tokens and $10 per million output tokens, with cached input tokens 95% off. A footnote says the introductory period ends on November 30, 2026, after which the public model costs $4 and $20.
Google says Argon ranks first on the Vals Index. That is their ranking. The post does not include the comparison table.
[1]The engineering numbers are Google's own
Google says that over the past few weeks, engineers using Argon in Antigravity increased agentic code submissions by more than 35% and reduced rollbacks by nearly a third. In another internal note, end-to-end inference of Gemma 4 E2B on one A100 through vLLM rose from about 37,900 tokens per second to more than 108,000. They call that 2.85 times, with matching tokens and no quality drop. These are Google's measurements.
On the embodied-reasoning benchmark ERQA, they report 87% and say that is higher than other frontier models. The post does not give the other scores.
[1]