Google recently unveiled its latest AI innovation, Gemini 1.5 Flash, claiming it to be 20% faster than OpenAI’s newest model, ChatGPT-4o. Available to the general public, this cutting-edge AI has already attracted prominent clients, including Moody’s and Uber Eats.
Initially introduced at Google I/O in May, Gemini 1.5 Flash has been in public preview for the past month, allowing users to test its capabilities and provide feedback. This new model boasts impressive processing power, capable of analyzing one hour of video, 11 hours of audio, or over 700,000 words in a single query. In a recent demonstration, Google showcased how the AI could analyze a 14-minute video in just one minute.
Additionally, Gemini 1.5 Flash is 40% faster than ChatGPT-3.5-turbo and can handle a greater volume of information, enabling users to include extensive context—whether audio, video, code, or text—in their queries and receive more accurate answers.
Google Cloud CEO Thomas Kurian highlighted the model’s market-leading performance, emphasizing its cost-effectiveness. The AI is available to developers in a limited capacity for free, with variable pricing based on data input requirements.
For those needing even more robust capabilities, Google offers Gemini 1.5 Pro, which is approximately ten times more expensive but can process 22 hours of audio and 1.5 million words. This allows for comprehensive analysis, such as examining an entire company’s decade-long financial history in a single query.
Google’s recent updates to Gemini 1.5 Flash, the flagship Gemini 1.5 Pro, and its image generator, Imagen 3, position its AI suite as the premier generative AI platform for enterprise applications. Notable customers include Uber Eats, Moody’s, and Shutterstock, who are leveraging Google’s AI advancements for enhanced decision-making processes.
A key feature driving this adoption is the enhanced “grounding” ability, which links answers to sources and provides a trustworthiness score. This feature is crucial for reducing errors and improving the factual accuracy of AI-generated responses. Moody’s Analytics Chief Product Officer Nick Reed emphasized the importance of grounding in using generative AI for decision-making, especially in financial services.
Looking ahead, Google plans to introduce a new industry-specific grounding tool in the third quarter. This tool will enable financial analysts to ground their AI queries with Moody’s data and legal experts to reference Thomson Reuters sources, further enhancing the AI’s reliability and applicability in various sectors.
