Google Opens Gemini 4 Argon to Developers, Undercuts Rivals on Price (For Now)

Two weeks after Google handed its newest frontier model to a small circle of cyber defenders, the company has thrown the doors open. On October 13, Alphabet's Google released Gemini 4 Argon to developers and businesses worldwide, seeking to regain ground against OpenAI and Anthropic after months of delays and a leadership shakeup at its DeepMind lab, according to reporting from Reuters. The pricing is the most aggressive part of the story: an introductory $2 per million input tokens and $10 per million output tokens, but only until November 30. After that, the rate doubles to exactly what Anthropic charges for Claude Opus 5.5.

What launched, and who gets it

Gemini 4 Argon is Google's flagship AI model, which the company says excels at "deep work" on complex, multi-step tasks across coding, finance, law, cybersecurity, and science. Access breaks down by customer type. Developers on the paid tier of the Gemini API and Google AI Studio get the model under the ID gemini-4-argon, now listed as stable in Google's API documentation. Google AI Pro and Ultra subscribers can use it in the Gemini app and in AI Mode in Search. Google's agentic coding tool, Antigravity, is available to Ultra subscribers, or to businesses through Gemini Enterprise.

Early enterprise testers named by Google include Deloitte, LexisNexis, Ambience Healthcare, and Isomorphic Labs, a cluster that points to where Google thinks the model's strengths lie: professional services, legal research, healthcare, and science. In a bid to seed the scientific community, Google said it would offer 10,000 scientists six months of free access to Antigravity's highest usage tier. Alongside the model release, the Gemini app gained a tasks mode that Google says can keep working "even when you're not at your computer," with connections to more than 40 apps including Instacart, OpenTable, Adobe, and Shopify.

A price war with an expiry date

Google's launch pricing is the clearest signal of its strategy. At $2 per million input tokens and $10 per million output tokens, with cached input tokens discounted 95 percent to $0.10 per million, Argon undercuts every major frontier rival at the entry point. Anthropic's Claude Opus 5.5 charges $4 and $20 per million input and output tokens. OpenAI's GPT-6 Astra charges $10 and $50. On a per-task basis, Mashable put Argon's cost at about $1.99 against the Artificial Analysis Intelligence Index, roughly 60 percent of what GPT-6 Astra costs for comparable work.

The $2 and $10 rates are a seven-week window, not a strategy. Anyone building on Argon should price their roadmap against the $4 and $20 reality that arrives December 1.

The catch is that the launch price has a hard end date. Reuters and VentureBeat report the introductory rates run only until November 30, 2026, after which Argon costs $4 and $20, identical to Claude Opus 5.5, and twice as much as Claude Sonnet 5.5 or GPT-6.1 Sol. That makes the pricing announcement as much a migration incentive as a genuine price war: Google gets seven weeks to pull developers and enterprises onto its platform, after which the market price reasserts itself.

The numbers Google is putting up

The launch came with a stack of performance claims, and it is worth separating Google's own numbers from independent ones. Independent scores first: Argon ranked first on the Vals Index at 68.90 percent as of October 12, ahead of Claude Sonnet 5.5 at 67.04 percent and Claude Opus 5.5 at 66.97 percent. On the Artificial Analysis Intelligence Index, it tied OpenAI's GPT-6 Astra and trailed two Anthropic models as of October 2, before the wide release.

Google's own figures, which the company disclosed in its announcement post, are more dramatic but not yet independently verified. The company says teams using Argon in Antigravity increased agentic code submissions by more than 35 percent while reducing code rollbacks by nearly a third, though it did not disclose the measurement period or the comparison baseline. The model carries a one million token unified input-output limit, with a new "continuation protocol" for longer work, and Google reported a 0.7 percent attack success rate against indirect prompt injection across 15 adaptive attempts in the Gray Swan evaluation. Treat the first-party numbers as directionally interesting and the independent benchmarks as the ones to watch.

Why Argon's arrival matters: the rocky road here

To understand why this launch matters, rewind a month. In September, Google announced it would no longer release Gemini 3.5 Pro, a model CEO Sundar Pichai had originally said would ship in June. Around the same time, the company overhauled DeepMind: co-founder and CEO Demis Hassabis stepped aside, and several leaders of the Gemini model team left the company. A flagship launch under those conditions is not routine. It is a credibility exercise, and the pricing suggests Google knows it has to earn developers back rather than assume they will wait.

The security dimension is notable too. Google said it will expand its Fairwind Program to hundreds of additional partners. Fairwind gives vetted cybersecurity defenders a version of Argon without its cyber guardrails, the same privileged access the model had exclusively in its first two weeks. The program reflects a growing industry pattern: the most capable models are dual-use by nature, and Google is trying to keep defenders ahead by giving them the unguarded version while everyone else gets the fenced one.

What this means for builders

The practical takeaway for developers is straightforward. If you were waiting for a public model ID to start evaluation work, that wait is over: gemini-4-argon is stable and documented. The seven-week launch pricing is a genuine, if brief, discount window worth using for evaluation and early integration. But build your cost models around the December price, not the October one, because a business case that depends on $2 and $10 stops being valid the moment November ends. Alphabet's shares moved only slightly on the news, up 0.4 percent to $354.10 in midday trading on October 13, a reminder that the market has seen flagship model launches before. Whether Argon changes Google's position in the enterprise AI race will depend less on launch-day pricing and more on whether the code submissions, the rollback rates, and the benchmarks hold up in the wild. (Analysis)