Gemini is trained to seamlessly understand and reason on all kinds of inputs, and this should make it pretty capable in the face of complex coding requests and even physics problems.
Gemini is being 'made' into three sizes: Nano, Pro and Ultra. Nano is on-device, and Pro will fold into Google's chatbot, Bard. Gemini Pro apparently outscored the earlier model, which initially powered ChatGPT, called GPT-3.5, on six of eight AI benchmarks. Gemini Ultra scored higher than any other model, including GPT-4, on some benchmark tests.
[
add
]
[
|
|
...
]