Google releases Gemini 4 Argon with cheaper introductory API pricing

Google has launched Gemini 4 Argon, its long-awaited frontier AI model, after what The Verge describes as months of lagging behind rival labs. The release matters for creators, marketers and developers because Google is pairing broader product access with aggressive introductory API pricing, while claiming gains in reasoning and complex work.

The company says Argon can compete with leading models from OpenAI and Anthropic. For users deciding where to run writing, coding, research or agent-style tasks, the key questions are availability, cost and whether Google’s claimed quality improvements show up in day-to-day work.

What Google says is new in Argon

Google says Gemini 4 Argon brings stronger reasoning and better performance on complex workflows. The company also says it has reduced some common chatbot annoyances, including hedging, repetition and formulaic phrasing.

In benchmark claims cited by The Verge, Google says Argon beats rival models including GPT-6 Astra, Claude Fable 5.1 and Claude Opus 5.5 across areas such as coding, cybersecurity, science, finance and law. As always with vendor benchmarks, teams should treat those claims as a starting point for testing rather than a substitute for trying the model on their own prompts, documents and workflows.

For people already using Google Gemini, the most relevant promise is not a single benchmark category but whether Argon handles multi-step tasks with fewer restarts: drafting and revising campaign copy, analyzing research, generating code, or moving through a structured brief without getting stuck in generic answers.

Who can use Gemini 4 Argon now

Argon is rolling out today to Google’s higher-tier Pro and Ultra subscribers in the Gemini app and in AI mode in Search, according to the report. Ultra subscribers also get access through Antigravity, Google’s agentic development platform.

Google has not said when, or whether, the model will reach Plus subscribers or free users. That makes the first wave most relevant to paying users who already rely on Gemini for work, and to developers on Google’s highest consumer tier who want to test agentic coding or automation ideas in Antigravity.

The rollout also followed a limited release to selected “trusted cyber defenders” for risk testing. That mirrors the pattern of other recent frontier model launches, where companies give a smaller group early access before opening the system to a broader audience.

For small teams, the practical takeaway is simple:

• If you are on Pro or Ultra, Argon is worth testing against your current default model.
• If you are on Plus or free access, there is no confirmed timeline in the source.
• If you build with the API, pricing may be the bigger immediate change than app access.

API pricing could shift model testing budgets

Google is launching Argon with an introductory API price of $2 per million input tokens and $10 per million output tokens. The Verge reports that this is roughly one-fifth the price of OpenAI’s GPT-6 Astra and Anthropic’s Fable 5 models.

Those introductory prices are set to double at the end of November, according to the report. Even after that increase, Argon is described as remaining cheaper than frontier rivals on a per-token basis.

That does not automatically mean every workflow will cost less. The source notes an important caveat: different models may use different token counts to complete the same task. A model with a lower per-token price can still be more expensive in practice if it generates longer outputs, needs more retries, or requires longer prompts to get reliable results.

For developers and agencies, the sensible move is to compare end-to-end task cost rather than headline token rates. Test the same set of prompts across Argon and your current model, then measure whether the answer quality, token use and number of follow-up prompts justify switching. Our ChatGPT vs Claude vs Google Gemini buying guide can help frame that evaluation across the main chatbot ecosystems.

Google’s frontier push is back in focus

Argon is Google’s first major frontier model release since Gemini 3 in late 2025, according to The Verge. The launch comes after a period in which models from OpenAI and Anthropic put pressure on Google’s position in the frontier AI market.

The company says it is already using Argon internally. Examples cited in the report include finding and fixing cybersecurity vulnerabilities, reducing the computing power needed for YouTube’s comments section by half, and debugging the drone physics simulator for Wing, Alphabet’s drone delivery company.

The timing also underlines how crowded the top tier of AI has become. Argon arrived a day after Meta announced its own frontier model, Muse Wonder Full; we covered that launch here: Meta launches Muse Wonder, its first frontier AI model family.

For creators and small businesses, this competition is useful only if it turns into better tools, clearer access and predictable costs. Argon gives Google a new high-end model to put inside Gemini, Search and developer workflows. The next test is whether users see enough practical improvement to make it their default for real work.