I think it's about convention and adoption - "first taste is free" to get as many ppl as possible accustomed (read: addicted) to using GenAI. Plus most majors are currently releasing open weight models.
...but i suspect few to none will continue releasing free and open models when they start cutting too deeply into token profits.
We'll likely see smaller open models, but few who have the resources to train large, frontier sized models will want to do so for free... maybe China, but they're also a dubious choice for many.
...but i suspect few to none will continue releasing free and open models when they start cutting too deeply into token profits.
We'll likely see smaller open models, but few who have the resources to train large, frontier sized models will want to do so for free... maybe China, but they're also a dubious choice for many.