• Source:JND
Add as a preferred

Gemini Pro AI: Search leader Google has released the 'Pro' iteration of its 'most capable' AI model for developers and enterprise customers with a diverse set of AI tools and associated infrastructure. The release follows the announcement of the model and release of the 'Nano' version of Gemini for Pixel 8 Pro users. This can be seen as an attempt to garner more developers to bolster the progress of Google's cloud services.

For developers, the model will be available via Gemini's Application Programming Interface (API) within Google AI studio, the tech giant's in-house web-based developer solution. Business customers can access Gemini Pro using the Google Cloud's enterprise platform Vertex AI. For further customisations, developers can move their AI Studio code into Vertex AI.

Also Read: Google Gemini AI Launched: AI Model Set To Beat GPT-4, Human Experts; Know All About Its Versions

According to the tech giant, the model is available in 38 languages and 180 countries across the globe. It can accept input in the form of texts and output results in textual form. For multi-modal based capabilities (currently images only), users can use a separate solution - Gemini Pro Vision. However, the results are limited to textual format.

At the time of the announcement a few days back, Google claimed that Gemini Pro defeats GPT-3.5 in six benchmarks out of eight. This includes massive multitask language understanding (MMLU) that deploys a set of 57 subjects ranging from physics, maths, law, history and more for analysing performance.Demo of Imagen 2 shared by Google. (Image:Google)

Gemini Pro and Pro Vision will be available for developers (up to 60 requests/minute) via the Google AI Studio. Vertex AI users can access the model without any extra costs until the model reaches general availability in 2024. Besides Gemini Pro, the company also introduced an upgraded version of Imagen 2.

Also Read: 'Once Upon A Time...': Google's Gemini AI May Narrate Your Life As A Story Tale Via Your Phone's Data; Here's How 

The image processing model offers improved image production and better text rendering across different languages and other features. These include creating logos, emblems, lettermarks and more. Support for Hindi, Korean, Chinese, Japanese, Spanish, and Portuguese is now available in preview.

Plus, the company also introduced MedLM, building on the company's Med-PaLM 2 model, AI models for the healthcare sector in the United States and a host of Duet Al-based enhancements ranging from Duet Al for Developers and Duet AI in Security Operations.