보관된 글 — US — techcrunch
Google launches Gemini 3 with new coding app and record benchmark scores
Google has launched Gemini 3, its most advanced language model, featuring significant improvements in reasoning capabilities and a new coding app called Google Antigravity. The model scores record benchmarks and comes amidst rapid advancements in the AI industry, positioned as a strong competitor against similar offerings from OpenAI and Anthropic.
핵심 정리
- Google launches Gemini 3, an advanced language model with superior reasoning capabilities.
- Gemini 3 achieved a record score of 37.4 on the Humanity’s Last Exam benchmark.
- The launch coincided with releases from OpenAI (GPT 5.1) and Anthropic (Sonnet 4.5), indicating a competitive landscape.
- The new coding interface, Google Antigravity, offers integrated coding support via a chat window, command-line, and browser.
- Gemini app boasts over 650 million monthly users and 13 million developers using the model.
왜 중요한가
- Advancements in AI, particularly LLMs (Large Language Models), reflect the ongoing technology race in 2025.
- The rapid development of AI tools like GPT-5 Pro and Sonnet 4.5 shows intensified competition in the AI landscape.
- Growth in developer interactions suggests a trend towards integrating AI functionalities into software development workflows.
With Gemini 3, Google raises the stakes in the AI model competition, showcasing remarkable advancements in reasoning and providing new tools for developers, reflecting the fast-paced evolution of AI technology.