Alphabet Inc.’s Google is preparing to unveil an upgraded version of its artificial intelligence framework, introducing Gemini 1.5 Pro. This advanced iteration is specifically designed to handle larger volumes of text and video, marking a significant progression in the field of generative AI.
The debut of Gemini 1.5 Pro, scheduled for Thursday, is aimed at cloud customers and developers, showcasing Google’s commitment to demonstrating its expertise in the rapidly evolving artificial intelligence sector.
Oriol Vinyals, Google’s Vice President and co-tech lead for Gemini, highlighted the model’s strong foundation in cutting-edge research. Vinyals expressed enthusiasm about the global community’s response to the new features during a briefing with journalists.
Google affirms that the mid-size Gemini 1.5 Pro model matches the performance level of the larger Gemini 1.0 Ultra model. This launch follows OpenAI’s success with ChatGPT in late 2022, positioning Google as a leading player in advanced generative AI technology capable of producing text, visuals, and videos based on user inputs.
Initially introduced by Google in December, Gemini offered three distinct versions tailored for various functions and compatible with a range of devices, from mobile devices to extensive data centers. In response to advancements by Microsoft Corp. and OpenAI, Google aims to attract users with more powerful tools.
Gemini 1.5 Pro distinguishes itself by enabling faster and more efficient training, along with the ability to handle large amounts of data based on user prompts. Google states that Gemini 1.5 Pro can process up to an hour of video, 11 hours of audio, or over 700,000 words in a document — achieving a milestone known as the “longest context window” among large-scale AI models. This surpasses the data processing capabilities of the latest AI models from OpenAI and Anthropic, according to Google.
In a prerecorded video demonstration, Google showcased the capabilities of Gemini 1.5 Pro. For example, the AI model analyzed a 402-page PDF transcript of the Apollo 11 moon landing and accurately identified quotes highlighting “three funny moments.” Another demonstration involved pinpointing a specific scene in a 44-minute Buster Keaton film based on a rough sketch provided to the model.
Despite its advanced features, Google acknowledges that Gemini 1.5 Pro, like all generative models, is not without flaws. The AI model may exhibit imperfections such as hallucinations, occasional performance issues, and challenges in understanding user intent, requiring diverse questioning techniques for accurate responses. Vinyals emphasized that the model is still in an experimental and research phase, with ongoing efforts to enhance its performance.
Developers can explore Gemini 1.5 Pro through Google’s AI Studio, while selected cloud clients can access the model in a private preview on the enterprise platform, Vertex AI. Additionally, Google announced the expanded availability of its large-scale Gemini 1.0 Ultra, offering access to a broader global customer base on Vertex AI.
