[Google Launches Gemini 4 to Compete with AI Rivals, Internal Opinions on Programming Performance Vary] According to a report by Jinse Finance, on October 1, Google began rolling out its highly anticipated flagship AI model, Gemini 4 Argon. However, within the company, there are doubts about the model's actual performance in critical areas such as programming. On Wednesday, Google made the model available to a small group of trusted cybersecurity partners and stated that it plans to expand its usage after further testing, initially offering it to paid users. The company claims that Gemini 4 achieved leading results in multiple benchmark tests, including outperforming OpenAI's Astra model in a test measuring security capabilities. However, some internal sources argue that these metrics do not provide the full picture. According to individuals directly familiar with the project, while Gemini 4 performed well in benchmark tests measuring model efficiency, its real-world performance during employee use was less impressive, particularly for certain programming tasks.