Henrik R
Gemini has been launched by Google.
According to Dagens.dk, Google has recently unveiled their latest language model, Gemini, which reportedly surpasses OpenAI’s ChatGPT in 30 out of 32 aspects. Over the past year, OpenAI’s ChatGPT has been leading the AI language model scene, but Google has now stepped up with the launch of Gemini.
This new AI model, as claimed by Google, outperforms OpenAI’s GPT-4 in most benchmarks for large language models (LLMs). Gemini has been integrated into Google’s Bard chatbot, marking a significant upgrade since the chatbot’s inception. Gemini is available in three versions: Nano, Pro, and Ultra, each tailored to different capacities and applications. Initially, consumers will have access to the Nano and Pro versions through various tools and interfaces, while the Ultra version is reserved for clients requiring advanced AI capabilities.
Interestingly, Google only reveals one benchmark where their model falls short against GPT-4, in the HellaSwag test, which assesses “commonsense reasoning for everyday activities.” Bard, powered by Gemini Pro, is now available in 170 countries in English, with plans to add more languages soon. Early next year, Bard will be upgraded to Gemini Ultra, which is still undergoing fine-tuning and security checks.