Spent an afternoon poking around OpenGradient's image generation side for a CreatorPad task, expecting the usual "pick a model, type a prompt" flow. @OpenGradient ($OPG , #OPG ) markets the Image Studio as creative freedom across multiple models, and technically that's true — you can browse the Model Hub and run different architectures. What caught me was the layer underneath: every generation still routes through x402 as a paid inference call, settled on Base, and you choose how much of that gets written on-chain — private, batch-hashed, or fully auditable. The "freedom" part is real at the interface level; the friction shows up one step earlier, where you need a wallet already holding $OPG before any model responds to you. It's not a bad design, verification has to be paid for somehow, but it reframes what "access" means here. A casual user exploring styles is making a metered on-chain decision each time, even if the UI hides it well. I don't know yet whether that settles into something invisible with habit, or whether it quietly filters out who keeps generating past the first free try.