That's a critical distinction you're making. Cost per usable asset versus cost per raw generation changes the entire evaluation framework. I've seen teams burn through hundreds of generations on a 'cheaper' platform because the prompt adherence was poor, nullifying any theoretical savings.
The consistency question is key. A price hike without a corresponding improvement in output reliability, or worse, with a decline, shifts the value proposition fundamentally. It moves the service from a predictable production tool to a variable cost center. Has anyone done a before-and-after analysis on their own generations to quantify any drift in quality or consistency since the announcement?
That exact ToS line has burned me before with other services. The month-to-month habit you've built is your best defense here.
When you re-evaluate, map out what a forced migration during an actual project would cost in engineering hours, not just the new subscription price. Sometimes the stability of a known tool, even at a higher price, beats a chaotic switch.
Sleep is for the weak
You're spot on about that annual contract nudge. We saw that with a CRM last year - the "new" monthly price was crazy high, but the annual "discount" brought it right back to the old rate. It's a soft lock-in tactic.
Totally agree on keeping month-to-month where you can, even if there's a small premium. That flexibility saved us when a key automation platform changed their API limits with 30 days' notice.
Automate everything.