Google launches ChatGPT rival
In response to the success of OpenAI's ChatGPT since its launch a year ago, Google has officially launched its latest large language model, Gemini, its “most capable” AI model ever, marking a significant leap to catch up in the AI race. Notably, Google's journey into large
In response to the success of OpenAI's ChatGPT since its launch a year ago, Google has officially launched its latest large language model, Gemini, its “most capable” AI model ever, marking a significant leap to catch up in the AI race.
Notably, Google's journey into large language models has been shaped by a sense of urgency – especially in the wake of OpenAI's ChatGPT dominating the AI conversation, as it intends to reestablish itself as the world leader in AI.
The newly launched model, Gemini, boasts multimodal capabilities allowing it to seamlessly work with images, and video rather than just text. With this, it marks an improvement over its previous AI system Bard at tasks such as summarizing content, brainstorming, writing and planning. Overall it is capable of more advanced reasoning and planning.
The model comes in three versions: Gemini Nano, Gemini Pro, and Gemini Ultra, and is expected to be integrated into various Google products, including search, ad products, and Chrome. Bard will be powered by Gemini Pro, while Pixel 8 Pro users will get a few new offline features thanks to Gemini Nano. On the other hand, the Ultra is designed for data centres and enterprise applications with plans for integration into Bard and availability through a cloud API in 2024.
But to answer the important question of how Google's new Gemini system compares to OpenAI’s GPT-4, the company says it undertook a meticulous analysis, subjecting both models to 32 benchmarks. The results, according to Google DeepMind CEO Demis Hassabis, show Gemini outscoring ChatGPT in 30 out of 32 benchmarks.
Meanwhile, it is still too early to tell how big of an improvement Gemini has over ChatGPT, or even if it has any at all since it has yet to go through public testing which will be the true test of its capabilities. It however does highlight the rapid and potentially concerning pace of development in the generative AI landscape.
The rollout of the Gemini model has already commenced within Google's chatbot Bard for the English language setting, with availability in more than 170 countries and territories.
