Artificial intelligence news.
Straight from the original announcements.

Vercel Adds Google’s Gemini 4 Hydrogen to AI Gateway

November 12, 2026By the AI News Bites Editorial TeamSource: Vercel
Vercel Adds Google’s Gemini 4 Hydrogen to AI Gateway
Image: Vercel

Google’s Gemini 4 Hydrogen is now available through Vercel’s AI Gateway, succeeding Gemini 3.8 Flash with improvements aimed at coding agents, knowledge work and tool use. Vercel says the model remains in the same latency class while handling longer engineering tasks in fewer agent steps.

Key points
  • Gemini 4 Hydrogen is now available on Vercel AI Gateway.
  • It succeeds Gemini 3.8 Flash and remains in the same latency class.
  • Vercel highlights longer engineering tasks, tool orchestration and multimodal reasoning.
  • Developers should use google/gemini-4-hydrogen as the model name.
  • AI Gateway has no markup on provider pricing and no platform fee on inference.

A model focused on longer agent tasks

Gemini 4 Hydrogen is designed to support longer engineering tasks and produce better-researched documents, spreadsheets and presentations with fewer agent steps, according to Vercel. The model also follows system instructions, formatting requirements and task boundaries more closely.

Vercel says Hydrogen provides more reliable tool orchestration across long conversations. This is aimed at workflows in which an AI model must plan work, call external tools and maintain context over an extended interaction.

Improved reasoning and grounding

The update includes improvements in visual reasoning and technical problem-solving, supporting multimodal workflows. Vercel also says stronger grounding reduces unsupported claims in search-backed answers and helps prevent invented tool or test results in coding reports.

These capabilities are presented as improvements over Gemini 3.8 Flash, which Gemini 4 Hydrogen succeeds. The source does not provide benchmark results or additional latency figures, but says the model stays in the same latency class.

How developers can access Gemini 4 Hydrogen

Developers using AI Gateway should specify google/gemini-4-hydrogen as the model name. Vercel provides an AI SDK example that streams a TypeScript retry helper with exponential backoff and jitter.

The model can also be used in Claude Code, Codex, Cursor and other agents. To set it up, users should install the latest Vercel CLI and run npx vercel ai-gateway setup, then select google/gemini-4-hydrogen in the agent. Vercel also directs users to its coding agents guide and model playground.

AI Gateway offers a unified API for calling models, tracking usage and cost, and configuring retries, failover and performance optimizations for higher-than-provider uptime. It also includes custom reporting, Zero Data Retention support, budgets for API keys and routing rules.

Vercel says AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including Bring Your Own Key requests. For users, the launch adds Gemini 4 Hydrogen to an existing interface for model access, monitoring and routing, while giving developers another option for long-running coding and knowledge-work agents.

Related news

Alibaba’s Qwen4-Max arrives on Vercel’s AI Gateway
ModelsNovember 15, 2026

Alibaba’s Qwen4-Max arrives on Vercel’s AI Gateway

Alibaba’s Qwen4-Max is now available through Vercel’s AI Gateway, giving developers access to a reasoning model with native vision and a 1M-token context window. Vercel…

Vercel adds Thinking Machines’ Inkling 2 Small to AI Gateway
ModelsNovember 12, 2026

Vercel adds Thinking Machines’ Inkling 2 Small to AI Gateway

Vercel has added Thinking Machines’ Inkling 2 Small to AI Gateway, bringing a multimodal reasoning model for coding agents, visual understanding and professional…

Vercel adds Google’s Gemini Decision 1 to AI Gateway
Products and toolsNovember 11, 2026

Vercel adds Google’s Gemini Decision 1 to AI Gateway

Vercel has made Google’s Gemini Decision 1 available through AI Gateway, giving developers a multimodal model for making decisions from text, images, audio, video and…