Lates News

Google launches its new-generation frontier model Gemini 4 Carbon, ranking first on the latest AI capability leaderboard, with particular strength in complex programming, long-duration autonomous task execution, and cybersecurity. Google says the model scores 88.2% on discovering software vulnerabilities, more than doubles the previous generation's performance in generating secure code, reduces the success rate of prompt injection attacks to 0%, and cuts high-severity security alignment issue alerts by nearly 80%. Pricing is the same as the previous generation: $2 per million input tokens and $10 per million output tokens, with cached input tokens priced 95% lower. It is initially being made available to select cybersecurity defenders, with broader access to follow gradually.