Google Bard Receives ‘Gemini Pro’ Support; Company Claims Most Significant Upgrade Since Initial Release
Google Bard now features Gemini Pro, Google’s latest multimodal language model (LLM), replacing PaLM 2 as its most powerful AI model to date.
The company claims this is the “biggest update” Bard has received since its launch in March 2023, allowing for more advanced reasoning, planning, understanding and more.
Gemini 1.0 is available in three different versions – Gemini Ultra, Gemini Pro and Gemini Nano. Google says the Bard chatbot uses a “fine-tuned” version of Gemini Pro, but this could be updated in the future.
Gemini Pro beats OpenAI’s GPT 3.5
According to the company, Gemini Pro, which now uses Bard, beats GPT 3.5, which supports ChatGPT, in comparisons. “On six of eight benchmarks, Gemini Pro outperformed GPT-3.5, including MMLU (Massive Multitask Language Understanding), one of the most important leading measurement standards for large AI models, and GSM8K, which measures elementary school mathematical reasoning,” notes Google. .
Additionally, a version of Google Bard powered by Gemini Pro is already available and is available in English in over 170 countries and regions. Expansion to different methods, new locations and languages is coming soon.
Bard ‘Advanced’ coming soon
The search giant has also confirmed that it will release Gemini Ultra-powered Bard Advanced next year. By doing so, Bard becomes even more powerful than the current version.
It’s not clear when the Gemini Ultra and Bard Ultra will be released, but Google is aiming for an early 2024 release. The company attributes this to “extensive security checks, including red-teaming by trusted third parties and further refinement of the model using fine-tuning and learning with human feedback (RLHF)” before Gemini Ultra and the applications it uses are released to the public.
