Investment firm Baird said that Google's Gemini generative AI chatbot ranked OpenAI's GPT-4 as the most advanced large language model. Analyst Colin Sebastian asked Google's chatbot to rank the most advanced models, with OpenAI's model topping the list, followed by Google's PaLM2, unveiled by the company at its I/O developer conference in May.
Four others in the top five were Anthropic's Claude2 and Meta(META.US)'s LLaMA2. (Meta released the latest version of its LLaMA3 model, Llama3.1405B, last month.)
Microsoft invested tens of billions of dollars in OpenAI in January, and the company's generative AI chatbot ChatGPT soon took the world by storm. Since then, OpenAI has rolled out several new models, including its GPT-4o in May, and last month, the company also released a smaller, more affordable version of GPT-4 called GPT-4omini.
Appfigures, a research firm, said that the GPT-4o model released in May also boosted OpenAI's revenue. The model can accept text, audio and images as input and generate text, audio and images as output. The GPT-4o model was unveiled on the first night of the Google I/O developer conference. The company updated its Gemini 1.5 Pro, introducing a smaller 1.5 Flash version of the Gemini model, and also added two new Gemma models. In the generative AI space, the competition is heating up, and the pace of AI development and investment is growing stronger.