Google's New AI Model: Faster, Cheaper, and Smarter

Google has just launched a new AI model called Gemini 3 Flash. This model is designed to be quick and affordable, making it a strong competitor in the AI market. It's now the default model in the Gemini app and Google's search AI mode.

Key Features and Improvements

• Performance Boost: A significant upgrade from Gemini 2.5 Flash, Gemini 3 Flash performs better on various tests, even matching top models like Gemini 3 Pro and GPT-5.2.
• Test Scores: Scored 33.7% on Humanity's Last Exam, a major improvement from the 11% scored by its predecessor.
• Multimodal Capabilities: Can process and understand videos, drawings, and audio recordings. Users can upload a video of themselves playing pickleball and ask for tips or draw a sketch and have the model guess what they're drawing.

Availability and Use Cases

• Default Model: Now the default in the Gemini app worldwide.
• User Choice: Users can still opt for the Pro model for complex tasks like math and coding.
• Enhanced Understanding: Better at understanding user queries and generating visual answers, including images and tables.

For Businesses and Developers

• Accessibility: Available through Vertex AI and Gemini Enterprise.
• Preview Mode: Accessible via API and Antigravity, Google's new coding tool.
• Specialized Tasks: Excels in video analysis, data extraction, and visual Q&A.
• Efficiency: Three times faster than the previous model and uses fewer tokens for certain tasks, potentially saving on costs.

Competition and Industry Impact

Despite fierce competition with OpenAI, Google has focused on continuous improvement and innovation in the AI industry. The launch of new models is pushing companies to stay active and competitive.