Google released Gemini 4 Argon, its latest frontier model, available today through the API in AI Studio and the Gemini platform, plus the Gemini app and AI Mode for subscribers.
The headline number for anyone paying per token: an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input discounted by 95%. Google says the model ranks first on the Vals Index, which measures performance across finance, coding, legal and tax work.
The engineering claims are specific rather than vague. On Chrome, Argon helped separate core webpage handling from a roughly 6-million-line legacy folder inside Chromium's 40-million-line codebase, completing a refactor across hundreds of files. Google says teams using Argon in Antigravity increased agentic code submissions by more than 35% while reducing rollbacks by nearly a third.
There are also speed claims on open-source work: researchers used it to make the xTB quantum chemistry package up to 2.3x faster at predicting molecular structures, and to get a 2.85x inference speedup on Gemma 4 E2B through vLLM.
Why it matters
- The price is the story. Frontier capability at $2/$10 with a 95% cache discount changes what is viable to run at volume.
- The Chrome refactor is the kind of task that normally takes a team months. If it holds up, that is a real shift in what one model can do to a large codebase.
- Benchmark leadership claims are worth treating carefully until independent runs exist, but the engineering detail is unusually concrete.