Google releases Gemini 4 Argon, API priced at $2 per million input tokens

Gemini 4 Argon is available globally today: developers and enterprises can call it via the AI Studio and Gemini platform APIs, Google AI Ultra subscribers can use it in Antigravity, and Ultra and Pro subscribers can access it in the Gemini app and Search AI Mode. Pricing is $2 per million input tokens and $10 per million output tokens, with a 95% discount on cached input tokens. Google says it ranks first on the Vals Index across economic-sector tasks including finance, coding, law, and tax. For developers using the Gemini API, Argon in Antigravity increases agent code commits by over 35% and cuts code rollbacks by nearly a third; the ADK team used it to auto-optimize google/adk-python overnight, reducing long-session memory preparation time by 40% and speeding cold imports by 21%; Gemma 4 E2B on A100 via vLLM achieves 2.85x end-to-end inference speedup, raising throughput from about 37,900 to over 108,000 tokens per second with no quality loss. Scientists and security teams are also using it: the xTB quantum chemistry package predicts molecular structures up to 2.3x faster and conformational searches 1.8x faster; Wiz used it to find and fix a public file-download bypass exposing a master key and granting admin privileges; Isomorphic Labs compressed complex PDF and patent analysis from weeks to hours.