Google DeepMind Releases Gemini 4 Argon, an AI Model Built for the Hardest Problems

At a glance

Google DeepMind has released Gemini 4 Argon, a new AI model designed to solve the most difficult problems. Developers, businesses, scientists, and general users can now use it. The company says it performs well on deep, multi-step work in complex real-world tasks such as coding, finance, law, cybersecurity, and science.

Key points

1. 1

   It ranked first in the Vals Index, an evaluation that compares multiple AI models.

2. 2

   API usage pricing starts at an introductory rate of $2 per million input tokens and $10 per million output tokens.

3. 3

   Google used it internally to clean up the Chrome browser code, reducing work that would have taken months to a few weeks.

4. 4

   It was used to make xTB, an open-source quantum chemistry tool, up to 2.3 times faster, and the changes have been merged into the official repository.

Why it matters

This news shows how deeply AI can take on complex work, and it could affect the performance of the AI services people use every day.