Google’s Gemini 4 Carbon goes to cyber defenders first

Models
Safety
Two weeks after Argon, Google showed a stronger model for security and long coding jobs, but only chosen partners may use it.
Published

October 28, 2026

What happened

  • On October 28, Google announced Gemini 4 Carbon, two weeks after Gemini 4 Argon. Google calls it its strongest model, built for cybersecurity (finding and fixing weak spots in software) and for long coding and engineering jobs.
  • Google says Carbon writes safe code more than twice as often as Argon on its internal tests.
  • Only a small group of security defenders can use it, through Google’s Fairwind Program. Google gave no date for a wider release. It said the price will be the same as Argon’s introductory price.
  • On October 30, Carbon was first on the Artificial Analysis Intelligence Index, with 59 points.

Why it matters for you

A model that is very good at finding holes in software can help attackers as much as defenders. So the labs now give their best cyber models to defenders first, as Anthropic did with Mythos in April. For most users this means the top-ranked model on a board may not be one you can buy. It also shows how fast the race moves: Google’s lead model changed twice in one month.

Sources

Notes for learning. Nothing here is investment advice. All news