Google finally releases Gemini 4 Argon

The Verge reports Google launched Gemini 4 Argon after months on the sidelines of the frontier race, claiming strong reasoning across complex workflows and fewer chatbot clichés. Google says Argon outperforms rivals including GPT-6 Astra, Claude Fable 5.1, and Claude Opus 5.5 on benchmarks spanning coding, cybersecurity, science, finance, and law.

Rollout hits Pro and Ultra subscribers in the Gemini app and AI Mode in Search, and Ultra on Antigravity; availability for Plus or free users is unclear. Introductory API pricing is $2 / $10 per million tokens—about a fifth the cost of some rival frontier list prices—doubling after late November but remaining comparatively cheap on a per-token basis.

The public launch follows limited release to trusted cyber defenders. The Verge notes the timing, a day after Meta’s Muse Wonder Full news, as evidence the frontier is not a two-horse race.

“Finally” is doing work in that headline, and it’s fair. Google’s silence between Gemini 3 and Argon invited a story of drift; shipping a model this broad is how you answer it.

We’re with The Verge on the competitive read: price is a weapon, access tiers are a bet on willingness-to-pay for depth, and the defender-first window is now table stakes for models with serious cyber skill. Claims that Argon beats named rivals across a buffet of benches should be treated as the opening brief, not the verdict.

The industry structure point matters more than any one score. Meta’s Muse and Google’s Argon landing near each other remind customers they have options beyond the OpenAI–Anthropic gravity well. Options only help if switching costs and eval harnesses keep up.

Google needed a credible frontier release. It has one on paper. Now it needs weeks of non-Google-authored receipts.

Frontier AI competition in 2026 includes OpenAI, Anthropic, Google, and Meta releasing highly capable models in close succession, with pricing and access tiers as much a battleground as raw benchmarks.