I pay for ChatGPT and Gemini, and while Sol is a total beast with anything text, it still poisoned my cucumber bed. Which I will be bitter about for at least a few years while the bed recovers. Gemini (even flash) is exceptionally talented at viewing photos and telling you what to do/what it is (and telling me I just misidentified the problem with my cucumbers and spraying off the "bugs" actually just spread the bacteria everywhere.)
I also had really poor results with all major models from plant identification to plant treatment which is weird considering how much training material is out there.
Reminds me of recent headline: Chinese farmer kills 25 acres of crops due to LLM pesticide recipe: https://www.tomshardware.com/tech-industry/artificial-intell...
However afterwards, still feeling odd that I didn't see many bugs, I double checked with Gemini who is my typical standby for vision tasks. Gemini pointed out that its actually the fungus Corynespora cassiicola...which spreads by water splashing, and can live in your bed for years once it gets into the soil.
So now having two answers I need to research and check myself, and sure enough, it was obvious match for the fungus and didn't look much at all like the mite damage.
There is another story a few days later of 5.6 looking at my heatpump install and telling me it was probably borked and to call an expert. Gemini told me to just open a valve and it was good.
I don't really trust any other LLM besides gemini for vision stuff. It also tracks as google is the only lab still pursuing vision related tasks.
And in a typical enterprise environment dealing with documents, images, and broader business reasoning skills matter. Agents are not just for code and text :)