Google's next-generation AI model, Gemini, has hit a snag, with its public launch pushed back to January 2024. The AI foundation model was supposed to have a big reveal this week.
2
Scenario
Gemini's release was delayed after Google "found the AI didn't reliably handle some non-English queries." CEO Sundar Pichai says they are focused on getting Gemini 1.0 out as soon as possible.
3
Outperforming GPT-4
Despite language struggles, insiders said Gemini has met or surpassed GPT-4's standards in some areas. It aims to outshine OpenAI's model with its advanced text and image generation capabilities.
4
Multimodal content generation
Google announced Gemini at I/O 2023, showcasing impressive multimodal abilities, previously not seen in other models. Gemini can understand text and images while efficiently integrating tools/APIs.
5
Performance
Gemini can handle multiple types of data. It is said to possess the capability to comprehend and generate text, images, and other forms of content from websites, sketches or written descriptions.
6
Google integration uncertain
It's unclear when Gemini will be incorporated into Google services like Bard, Search, and Workspace. However, Google plans to offer various sizes of Gemini, including "Gecko" for mobile devices.