(C) aiarabai.com

On December 18, Google announced the official launch of "Gemini 3 Flash," a groundbreaking AI model designed to deliver frontier-level intelligence at unprecedented speeds. This release marks a significant milestone in Google’s strategy to eliminate the trade-off between high-performance reasoning and operational efficiency.

According to Google, Gemini 3 Flash achieved a 33.7% accuracy rate on the "Humanity’s Last Exam," a rigorous benchmark designed to test expert-level reasoning. This performance is a massive leap from its predecessor, Gemini 2.5 Flash, which scored 11.0%. Most impressively, it nearly matches the performance of the high-end Gemini 3 Pro (37.5%), proving that "Flash" models are no longer "lite" versions but powerful engines in their own right.

The new model is now the default engine for the Gemini app, replacing the 2.5 version. Free users can immediately benefit from enhanced multimodal capabilities—processing text, images, and video simultaneously—and improved logical inference. Furthermore, Gemini 3 Flash has been integrated into Google Search's "AI Overviews," enabling more precise and rapid responses to complex queries.

Josh Woodward, Vice President of Google Labs, stated, "Historically, users had to choose between slow, expensive frontier models or fast, less capable ones. Gemini 3 Flash ends this compromise, offering the best of both worlds." Technical reports suggest the model features an expanded context window, allowing users to upload massive documents or hour-long videos for instant analysis. This launch positions Google to lead the increasingly competitive AI market by providing accessible, high-speed, and high-intelligence solutions for both consumers and developers.

[ⓒ Global Economic Times. 무단전재-재배포 금지]