Google Launches Gemini 4 Hydrogen at $1.50 Per Million Input Tokens

Google launched Gemini 4 Hydrogen on November 12, 2026, priced at $1.50 per million input tokens and $7.50 per million output tokens, in a global rollout spanning its developer, enterprise and consumer products. The model is adapted from Gemini 4 Carbon and occupies a specific role in the Gemini 4 lineup: where Carbon and Argon handle deeper architectural judgment and extended reasoning, Hydrogen targets rapid execution — writing code, running tests, verifying changes, research and tool-using agents. It accepts text, images, audio and video, and carries a March 2026 knowledge cutoff that Google says may be less current in some domains.

The launch matters because it formalizes a division of labor that the industry has been converging on: reserving frontier planning models for the hard stages of a task while a cheaper, faster execution model does the bulk of the work. At Hydrogen's price point, that arrangement now extends across Google Search, the Gemini app, Gemini for business, Google AI Studio, Antigravity and Gemini in Sheets for eligible subscriptions.

What does the new Orchestrator in Antigravity change?

The Orchestrator option in Google Antigravity, Google's agentic development environment, splits planning from execution within a single session. Carbon explores dependencies and produces a plan in a read-only phase. Hydrogen then edits code, updates packages and runs tests. If execution encounters friction, control returns to Carbon to inspect the state and repair the plan.

The design differs from a static router. A static router commits an entire task to one model based on the opening prompt. Orchestrator can change models as the work changes.

The Antigravity team reports near-Carbon task success on Terminal-Bench 4.1, a benchmark for terminal-based agentic coding, while using 24% less quota and finishing 19% faster on average. For parallel work, Carbon coordinates while Hydrogen powers worker and verifier agents in isolated workspaces.

Hydrogen is rolling out to paid Google AI Pro and Ultra subscribers in Antigravity. Orchestrator initially rolls out to AI Ultra subscribers, with enterprise access to both available through Gemini for business.

How fast is Hydrogen compared to earlier models?

Google's reported results position Hydrogen well ahead of Gemini 3.8 Flash, its previous fast-tier model:

• Hydrogen finishes tasks 2.5 times faster at the high thinking level than Gemini 3.8 Flash.
• It scores 57% on Terminal-Bench 4.1, versus 18.2% for Gemini 3.8 Flash.
• It cuts fabricated claims by more than fivefold compared to Gemini 3.8 Flash.
• On the Artificial Analysis Intelligence Index, a composite benchmark across model capabilities, Hydrogen scores 51 at the high thinking level — above Argon's 50 — and rises to 54 at the new xhigh level.

The model card contains a material qualification. Its cheating-adjusted Terminal-Bench 4.1 table lists Anthropic's Claude Sonnet 5.5 at 63.9%, above Hydrogen's 57.0%. Hydrogen improves sharply on Google's previous Flash model without leading every competitor in the same coding comparison.

What are the pricing and token-window mechanics?

Beyond the flat paid pricing of $1.50 per million input tokens and $7.50 per million output tokens, the Gemini API documentation lists:

• A rate-limited free tier.
• Cached input at $0.075 per million tokens through implicit caching, a mechanism that reuses identical prompt prefixes at reduced cost.

Hydrogen's input, thinking and output share one 1,048,576-token context window. Synchronous responses support up to 65,536 tokens per hop, meaning per model call in an agentic chain. Background execution or client-side continuation can extend generation up to the remaining capacity in that window.

The model offers four thinking levels: low, medium, high by default, and the new xhigh. Google recommends high for most complex tasks and reserves xhigh for harder multistep problems where high falls short. The company notes xhigh can overthink simpler prompts. These controls make the shared token budget consequential: reasoning consumes capacity that could otherwise hold input or final output, and higher settings also increase latency.

Where does Hydrogen appear outside the API?

The November 12 rollout covers Gemini for business, the Gemini app, Google Search, Antigravity and Gemini in Sheets for eligible subscriptions, in addition to the API and AI Studio. Free Gemini app users receive Hydrogen through automatic model selection when Google considers it appropriate for a prompt.

Google attributes several product-level improvements to the model:

• The Gemini app follows tone and formatting instructions more closely, draws on uploads and connected services including Gmail, Drive, Airtable and Dropbox, and selects tools more reliably.
• Google Search gains better handling of multipart or typo-filled queries, more visual responses and expanded query fan-out, a technique that issues multiple query variations to retrieve a wider range of links.

What do the safety assessments say?

Google reports reductions of up to 81% in agentic overreach — a model taking actions beyond its mandate — and 79% in under-service, a model declining actions it should perform, versus Gemini 3.8 Flash. The company also reports greater resistance to indirect prompt injection, an attack where hidden instructions in retrieved content steer the model. Those measures address failure modes that matter when a model can select tools and act across connected services.

The model card records one regression: automated comparisons show a 5.57-percentage-point increase in unjustified refusals versus Gemini 3.8 Flash. Its listed limitations include hallucinations, occasional slowness or timeouts, and potentially higher token use at elevated thinking levels.

The frontier-safety assessment in the model card found no material general-capability increase over Argon. Targeted tests placed Hydrogen below the biological-information and cybersecurity critical capability levels, which mark severe-scale risk in Google's safety framework, while reaching the alert thresholds in both domains. The card says fuller safety documentation is forthcoming.

Why the split matters

Hydrogen's broader significance is the division of labor it enables: frontier planning reserved for harder stages while a faster execution model handles work across widely used products. Whether that arrangement preserves the reported efficiency gains in everyday use — alongside the documented refusal and reliability limitations — is the central question for this rollout.