Google made Gemini 4 Carbon broadly available on 10 November 2026. The company describes it as its strongest model to date, aimed at advanced reasoning, complex coding, scientific research and multi-step agent workflows that verify their own work.
The model is now open to developers, business customers and Google AI Ultra subscribers. Before the wider release, Google ran a two-week preview limited to cybersecurity defenders, giving security teams early access ahead of general users.
Why it matters for IT teams
The defender-first rollout is a sign of how seriously vendors now treat the security impact of frontier models. For organisations already on Google Cloud or Workspace, Gemini 4 Carbon is the new top option to benchmark for coding assistants and agent pipelines.

Leave a Reply Cancel reply