Google parent company Alphabet (GOOGL) has launched “Gemini 4 Argon,” its newest and most powerful artificial intelligence (A.I.) model yet.
In releasing the new model, Alphabet said that it offers major improvements in coding, cybersecurity, and complex professional work.
The company said Gemini 4 Argon set a record in real-world software engineering, tied for first in cybersecurity, and leads performance in finance, legal, and other professional tasks.
Argon is being used internally at Alphabet to optimize memory at its data centres, freeing up hundreds of terabytes of memory without buying additional hardware, said the company.
The new A.I. model can also help with the development of quantum computers.
Alphabet said that it plans to rollout the new model in phases, starting with trusted cybersecurity partners while working with the U.S. government on pre-release safety evaluations.
Gemini 4 Argon is being launched nearly a year after Alphabet’s previous Gemini 3 model made its debut to strong reviews from analysts and users.
Alphabet said in announcing Argon that it’s looking to implement safeguards in four key cybersecurity areas to ensure the new model is secure.
Earlier this week, Alphabet CEO Sundar Pichai signed a voluntary agreement with U.S. President Donald Trump that aims to address A.I. safety concerns.
GOOGL stock has risen 40% over the past year to trade at $344.08 U.S. per share.
