@Quinnypig Realistically, I'm not sure there's enough of a day to day difference across a wide range of tasks to make a jump.
Claude, Gemini, and ChatGPT may each be better at some tasks, so I keep telling folks here to that it may be worth their time to play with all three in their heavily used niches.
Gemini, for example, generated a far better (people's choice winning!) chili recipe for our IT department's annual chili contest. ChatGPT gave me boring midwest chili. Thumbs up for that use...
@Quinnypig I concur with @davidseidl. Of the "free" single model web chat services, using the same prompt on all three helps illuminate the (in)consistencies of each and increases the probability of a useful answer.
On mobile, I like Quora's Poe app for the consistent interface to a very wide range of models ("bots") including all the major "free" and many or most of the subscription models. The matching web interface at https://poe.com makes it easy to move between mobile and desktop.
@Quinnypig Try Microsoft's Phi-3.
Axios VP is claiming you should be using it for very very high stakes applications. 😂
https://www.axios.com/2024/04/23/microsoft-open-source-small-language-model-phi