Google Unveils Gemini 1.5 Pro With 1 Million Token Context Window
Google announced Gemini 1.5 Pro, a new flagship AI model, on February 15, 2024, featuring a groundbreaking 1 million token context window. This expansion represents a tenfold increase over the 100,000 token window of its predecessor, Gemini 1.0 Pro. The larger context window allows the model to process and analyze significantly longer documents, codebases, and video content, enabling more comprehensive understanding and reasoning capabilities.
During a demonstration, Google showcased Gemini 1.5 Pro's ability to analyze a 402-page document, a 11-hour long video, and 30,000 lines of code within seconds. This capability is attributed to a new Mixture-of-Experts (MoE) architecture, which Google states makes the model more efficient and performant. The MoE architecture allows the model to selectively activate different parts of its neural network for specific tasks, leading to faster processing and reduced computational cost compared to traditional dense models.
While the 1 million token context window will be available to select developers and enterprise customers through a private preview starting December 2024, Google also announced that a 128,000 token version will be made available to all developers via the Gemini API in February 2024. This tiered approach aims to balance cutting-edge capabilities with broader accessibility. The company emphasized that the model maintains strong performance across various benchmarks, including reasoning, coding, and multimodal understanding, despite the significant increase in context length.
Google highlighted that Gemini 1.5 Pro's advancements are a direct result of ongoing research and development in AI, particularly in the area of large context windows and efficient model architectures. The company plans to continue iterating on the model, with future updates expected to further enhance its capabilities and expand its availability to a wider range of applications and users. The introduction of Gemini 1.5 Pro positions Google at the forefront of AI development, offering powerful tools for complex data analysis and understanding.
Original source — read the full reporting at the publisher:
Read on The Economist