Key points
- Argon is live now in the Gemini app and AI Mode for Pro and Ultra subscribers; Antigravity and AI Studio access is limited to Ultra, per Ars Technica.
- API rates are a temporary promotion: input costs $2 and output $10 per million tokens, doubling after November 30, according to Ars Technica and The Verge.
- Google says Argon ranks first on the Vals Index, which covers finance, coding, legal and tax work; the claim comes from Google's own announcement.
- Google is offering 10,000 scientists free top-tier Antigravity use for six months, with applications open and selection required.
Why it matters for marketers: Teams budgeting for AI content and workflow automation get a cheaper frontier option, but the promotional rates end November 30 and agentic access depends on plan tier.
Google has begun rolling out Gemini 4 Argon, its latest frontier model, after teasing it a few weeks ago and testing it first with a limited group of trusted cyber defenders. Google says the model leads other frontier systems on a range of benchmarks, including the Vals Index, which tracks work in finance, coding, legal and tax.
Access depends on what you pay. Pro and Ultra subscribers get Argon in the Gemini app and AI Mode in Search. Ars Technica reports that Antigravity and AI Studio are limited to Ultra plans, and that Google has not said when other tiers will get it. Free users are left out for now. Developers can call the model through the paid API.
For marketers, the pricing is the practical detail. The Verge reports the launch rates are around a fifth of what OpenAI and Anthropic charge for comparable models, but both outlets note they are introductory. Ars Technica says the rates double on November 30. The Verge adds that cheaper tokens do not always mean cheaper results, since models use different token counts for the same task.
Google also points to internal results, such as cutting server load for YouTube comments and finding old bugs in a drone simulator. Those are company claims, and none are marketing-specific. The more relevant signal is Argon's pitch on writing: Google says it hedges less, repeats itself less and follows style instructions better than earlier Gemini models.
Teams planning agent workflows should test Argon on real content and reporting tasks, then model costs at the post-November rates before committing.
Sources
This brief is our own summary. Read the full reporting at the source.
- Google DeepMind: Gemini 4 Argon: Built to solve your hardest problems
- Google AI: Gemini 4 Argon: Built to solve your hardest problems
- The Verge: Google finally releases Gemini 4 Argon
- Ars Technica: Gemini 4 Argon begins rolling out in Gemini app and Antigravity, but only if you pay
- Search Engine Journal: Google Adds Gemini 4 Argon To AI Mode For Pro & Ultra Subscribers via @sejournal, @MattGSouthern
- The Decoder: Gemini 4 Argon is now available worldwide as Google's most capable frontier model
- Latent Space: [AINews] AI Engineer NYC: Gemini 4 Argon, Vals AI Benchmarks
- MarkTechPost: Google Releases Gemini 4 Argon API With 1M Token Output Limit
