The company headed by Sundar Pichai steps on the accelerator in the semiconductor race. This Wednesday, Google presented ‘Gemini’, the model “more general and capable“that the firm has built “ever”, according to a statement issued by the company itself. As the text indicates, this new Artificial Intelligence model has been created with the aim of working with multiple types of information, “including text, code, image and video”.
This movement comes after Google was accused by the US Department of Justice of having hidden its advances in AI, despite being at the forefront in this matter, with the aim of not damaging its search engine, which exercises a dominance in the online search market.
According to the company, this new type of ‘large language model’ (LLM) will include three types of versions. On the one hand, Gemini Ultra, “their most capable category,” which shows signs of “more complex reasoning skills.” Specifically, this version has excelled in the benchmark MMLU (‘massive multitask language understanding’) tests, which use tests on 57 topics, including history, law, medicine, mathematics, physics and ethics.
On the other hand, Gemini Pro, a version that will be integrated into Bard, Google’s chatbot, which will initially be available in its English version in more than 170 countries. Finally, Gemini Nanowith the improved ‘summarize’ and ‘recorder’ abilities and which will be available on the Pixel 8 Pro smartphone. In addition, Android developers will be able to work with this latest version in their new models.
This movement comes nine months after, in March, OpenAI will release ChatGPT 4, the language model with which Gemini is intended to compete. On the other hand, this announcement comes weeks after Microsoft presented magnificent results in the third quarter, especially in its cloud business. In fact, Satya Nadella’s firm recently unveiled two new chips aimed at improving the ‘cloud’ branch of the business and reducing dependence on NVIDIA, the largest supplier in the market.
After these episodes, Google could not sit idly by. As a result, it has presented Gemini, a language model with which the firm wants to advance positions in the fight for Artificial intelligence. It remains to be seen whether, in the future, the Alphabet firm will choose to include these types of improvements in its search engine, since according to some experts, AI can put an end to Google’s search model.