Gemini 4 Hydrogen
In brief
Google's new Gemini 4 Hydrogen model features a 1M-token unified context window, four thinking levels, and long decode for agentic workflows.
Benchmarks and interface
11 from 2 sources
BenchmarkChart
BenchmarkChart
BenchmarkChart
BenchmarkChart
BenchmarkChart
BenchmarkTable
InterfaceScreenshot
InterfaceScreenshot
BenchmarkScreenshot
InterfaceScreenshot
InterfaceScreenshotHow it works
Developers call the model via the Interactions API or Google AI Studio. It processes multimodal inputs, supports long decode up to 1M tokens, and offers configurable thinking levels like xhigh for deep reasoning.
Developers building coding and agentic workflows
Availability
Sources
10 merged into this card- 1Put Gemini to work for youAnnouncement · blog.google · 17.11
- 2Introducing Gemini 4 HydrogenAnnouncement · blog.google · 12.11
- 3Gemini 4 Hydrogen: everyday coding and agentic workflows for developersAnnouncement · blog.google · 12.11
- 4Introducing Gemini 4 HydrogenAnnouncement · deepmind.google · 12.11
- 5Antigravity agentHow-to guide · aistudio.google.com
- 6Gemini 4 HydrogenHow-to guide · aistudio.google.com
- 7Gemini thinkingHow-to guide · aistudio.google.com
- 8What's new in Gemini 4 HydrogenHow-to guide · aistudio.google.com
More from Google
Gemini CLI adds standalone theme.json support
Google's gemini-cli v0.69.0 adds support for a standalone theme.json configuration file. Users can now define custom themes in a dedicated file.
Gemini CLI /stats heatmap subcommand
Google added a /stats heatmap subcommand to Gemini CLI v0.69.0 to track daily token usage. It lets developers monitor their token consumption directly from the command line.
