A feature that no one is talking about but which is even cooler: You can use the same technology to a) directly simulate color blindness (drive the L and M cones with the same signal) and b) in theory drive subsets of a dichromat's L cones with derived L- and M- cone amplitudes - does that make them sort of trichromatic? Such an amazing technology! The whole AO story ultimately links back to the Starfire Optical Range : https://en.wikipedia.org/wiki/Starfire_Optical_Range
If you only use pro, with a 7/1 ratio and no discounts or penalties the $6 plan gets you a total of 12M tokens. This assumes zero cached tokens thought.
I've found these kind of models (Deepseek + Xiaomi) to be absolutely excellent when it comes to writing documentation for code. We have a bunch of internal tasks that need to be documented for a non-technical team.
I added 20 USD in credits for the Xiaomi models a while ago and they've been happily writing and updating hundreds if not thousand of pages and I still have 7 USD left!
Through the magic of liner algebra it turns out that you can stimulate cones independently even with normal displays. Search for 'silent substitution'!