Google releases advanced AI models

Google has launched its new AI model, "Gemini 4 Aragon," which focuses on executing complex, multi-step tasks, especially in fields such as programming, software engineering, and cognitive work like finance and law. The model features an increased output limit of one million tokens, allowing it to handle large-scale projects and lengthy tasks without losing information. It demonstrates outstanding performance in software engineering and cybersecurity tests, including automatic discovery and recovery of vulnerabilities. Additionally, it has the ability to understand and analyze long video clips. The model is being gradually rolled out to laboratories and cybersecurity officials as part of the "Furwind" program, with strict protective measures to ensure security and responsible use. Its goal is to develop models that offer greater capabilities in executing complex work and multi-stage processes, with diverse results across various tests.

Size: 120%Text Size

Notice: This Is an AI-Generated Summary

Display The Full Article