Олександр КузьменкоAI Eng
14 October 2026, 13:14
2026-10-14
Google launches Gemini 4 Argon, a new AI model for complex codebases
Google has announced the release of Gemini 4 Argon, a new artificial intelligence model designed to perform multi-stage computational, engineering, and analytical tasks. The technology is aimed at developers, scientists, and the enterprise sector.
Google has announced the release of Gemini 4 Argon, a new artificial intelligence model designed to perform multi-stage computational, engineering, and analytical tasks. The technology is aimed at developers, scientists, and the enterprise sector.
This is stated in the company’s official announcement. The model is already available via the API in AI Studio, on the Gemini platform, as well as for Google AI Ultra and Pro subscribers in the Gemini app and in AI Mode in Search.
The developers note that Argon outperforms other leading models in comparative tests and ranks first in the Vals Index, which measures the effectiveness of AI in critical economic sectors such as finance, programming, law, and taxation.
For developers, the API usage cost is $2 per million input tokens and $10 per million output tokens. There is a 95% discount for cached input tokens. The model has a limit of 1 million unified I/O tokens.
According to Google, internal engineering teams are already using Argon to accelerate development. In particular, in the Antigravity environment, the model has helped increase the amount of code submitted by over 35% and reduce the number of code rollbacks by almost a third.
Other examples of use include:
Chrome browser refactoring: Argon helped separate deprecated features from the Chromium codebase, rewriting hundreds of files in a few weeks instead of a few months.
Optimization of the Agent Development Kit (ADK) framework on Python: running autonomously overnight, the model reduced the time for preparing sessions in memory by 40% and accelerated "cold" imports by 21%.
Accelerating quantum chemistry: Researchers make xTB package 2.3 times faster at predicting molecular structures.
Open model optimization: Gemma model inference speed increased by 2.85 times on A100 accelerators.
The company also opened applications for scientists to receive six months of free access to the highest Google Antigravity plan and launched the Fairwind program for cybersecurity professionals.
Як нейромережі бачать вільну та незалежну Україну? Тест dev.ua
Нейронні мережі для генерації зображень бачать світ по-своєму, їхню логіку зрозуміти часом зовсім неможливо. Але таки хочеться. На честь Дня Незалежності України редакція dev.ua вирішила провести невеликий експеримент.
Ми задали чотирьом різним нейронним мережам п’ять однакових запитів: «прапор України», «День Незалежності України», «український Крим», «перемога України» та «українці». Отриманими результатами ми ділимося з вами нижче.