Technology · 21 July 2026

Google launches cheaper AI model to find security flaws faster

Google has released a new AI security model, Gemini 3.5 Flash Cyber, designed to quickly identify and fix software vulnerabilities at a lower cost. This model is a more affordable option than larger, more expensive AI systems like Anthropic's Mythos. It will first be available to governments and trusted partners through CodeMender, Google's coding agent, and found 55 unique issues in the V8 JavaScript Engine.

By BriefTea Editorial · 21 July 2026 at 17:16

Reported by theverge.com · How we write briefs

Get the app

Understand this story

How do AI models get cheaper? How can it be faster AND cheaper?

More briefs

GeneralLondon temperatures falling to 13C by Sunday as wet, windy weather arrives PoliticsUK could get 'Switzerland plus' EU deal, says senior MEP SportsChelsea Captain Reece James Has Knee Surgery PoliticsHealey Rejects Mansion Tax Change That Would Double Homes Affected HealthGreat Ormond Street's Same-Day Test Finds Hidden Infections in Sick Children PoliticsUnion Warns 450 Jobs at Risk at UK's Biggest Gas Store Browse all stories →