Gemini 1.0 comes in three iterations: Ultra, Pro, and Nano, each tailored for specific tasks, with each modelled around its capacity to efficiently operate on different platforms, ranging from data centre’s to mobile devices.

Gemini’s introduction was accompanied by a demonstration of its ability to do stuff better than humans. Notably, Google claims that Gemini Ultra outperforms human experts on the massive multitask language understanding benchmark, showcasing its proficiency in tasks spanning math, physics, history, law and medicine.

In the realm of multimodal tasks, Gemini exhibits excellence, surpassing previous state-of-the-art models. Its native multimodality and advanced reasoning capabilities position it as a promising contender for applications in various fields, from scientific research to financial analysis.

Versatility in Coding and Beyond

Beyond its prowess in understanding and processing text and images, Gemini is also a dab hand at coding, it would seem. The model can comprehend, explain, and generate code in popular programming languages like Python, Java, C++, and Go.

Gemini’s coding capabilities extend to advanced systems, as demonstrated by the development of AlphaCode 2, a more sophisticated code generation system. This system excels in solving competitive programming problems that involve complex mathematical and theoretical computer science concepts.

How this ends up playing out in practice when it’s in the hands of devs, we shall see.

In its unveiling, Google was quick to emphasise its commitment to responsible AI development, incorporating extensive safety evaluations into the Gemini model. This includes assessments for potential biases and toxicity. The company engages with external experts and partners to stress-test models across various issues, addressing safety concerns ‘proactively.’

Recommended reading

• Google begins UK and UK roll out of ChatGPT competitor, Bard
• Robots will not replace human staff, says Amazon robotics chief
• Google’s DeepMind AI being used to train ML in nuclear fusion reactor

Gemini undergoes comprehensive safety evaluations, including assessments for potential biases and toxicity. The company collaborates with external experts to stress-test models across a range of issues, ensuring responsible development and deployment.

Gemini 1.0 is set to be integrated into various Google products, starting with Bard for advanced reasoning and planning. The model’s deployment extends to Pixel 8 Pro, becoming the first smartphone engineered to run Gemini Nano. Developers and enterprise customers gain access to Gemini Pro through the Gemini API in Google AI Studio or Google Cloud Vertex AI, marking a significant step forward in AI capabilities.