Google announced Gemini 4 Hydrogen on November 12, 2026, beginning a global rollout across Search, the Gemini app, AI Studio, the Gemini API, Gemini for business and Google Antigravity. The company positions the model for everyday coding and workflows in which AI carries out tasks using tools. Access conditions differ across those products. (blog.google)
Ars Technica reports that Hydrogen replaces Gemini 3.8 Flash across Google’s products. Google itself describes Hydrogen as an upgrade for teams migrating from Flash, but that language does not establish an API retirement deadline or mean existing applications automatically switch models. The distinction matters for developers planning changes to deployed software. (arstechnica.com)
Where Hydrogen is available
Google says Hydrogen is available in Search for all users globally. The Verge reports that Search product chief Robby Stein described it as AI Mode’s new default model. That does not mean Hydrogen produces every AI response across every Search feature, and the rollout announcement should not be read as confirmation that every account changed simultaneously. (blog.google)
In the Gemini app, Google announces Hydrogen in the model dropdown, but specifies a different arrangement for free users: they receive it through automatic model selection when appropriate for their prompt. The app’s help documentation corroborates that distinction. Free access therefore is not the same as unrestricted manual selection of Hydrogen for every conversation. (blog.google)
For builders, Google announces availability across all AI Studio tiers and through the Gemini API. Gemini for business is also included in the rollout. However, the announcement’s broad distribution list does not establish identical subscription entitlements, usage allowances or selection options across consumer, developer and business services. (blog.google)
Antigravity requires particular caution. Google’s launch announcement specifies paid Google AI Pro or Ultra subscriptions for consumer access. Separately checked Antigravity documentation conflicts with that wording: its model table lists Hydrogen for Free and Google AI Plus users as well, and its plans page includes Hydrogen among baseline models. These pages do not provide a consistent answer about consumer eligibility. (blog.google)
What Google says has improved
In Search, Google says Hydrogen better understands questions with multiple parts, difficult typos and highly specific requirements. It also claims improvements to visual answers and to query fan-out—the process of conducting related searches—which it says will surface a broader range of relevant links. Those are descriptions of intended improvements, not independently established gains in answer quality. (blog.google)
The distinction is important because understanding a complicated request and answering it correctly are separate tests. Google’s announcement presents examples and performance claims, but the reporting reviewed for this article does not independently demonstrate how reliably Hydrogen handles representative everyday searches. A more structured answer or a wider selection of links should not, on its own, be treated as proof of accuracy. (blog.google)
For developer workloads, Google claims task completion is 2.5 times faster at Hydrogen’s default high thinking setting than with Gemini 3.8 Flash. It also reports more than a fivefold reduction in fabricated claims. These remain company-attributed comparisons; this article does not independently reproduce the evaluations or establish equivalent gains in production applications. (blog.google)
Developer costs and migration boundaries
Google’s announced pricing is $1.50 per million input tokens and $7.50 per million output tokens, flat across the model’s one-million-token context window. It describes four thinking levels—low, medium, high and xhigh—with high as the default. These settings provide different reasoning options, rather than a single configuration underlying every possible performance comparison. (blog.google)
The practical distinction for existing Flash users is between a product replacement and an application migration. Google’s announcement addresses teams moving from Flash, while its API release notes separately confirm Hydrogen’s release. Neither the reported replacement nor the announcement alone establishes that an existing application’s chosen model has changed without developer action. (blog.google)
Distribution extends beyond Google’s interfaces
The release is also reaching third-party developer infrastructure. Vercel announced Hydrogen availability through AI Gateway on November 12. Vercel describes that service as a unified API for calling models, tracking usage and cost, and configuring retries and failover. That gives developers another distribution route, although availability through an intermediary is not independent evidence of better model performance. (vercel.com)
For users, the immediate change is a new model reaching several familiar services under different access arrangements. For application teams, the unresolved distinction is operational: broad product availability does not settle migration behavior or prove workload-specific improvements. The rollout is documented; the claimed gains still need evaluation, and Antigravity’s conflicting access guidance remains a qualification to any simple availability chart. (blog.google)
