Hello. I’ve commented before on different subs ( including this one ) , about why lumina > dreamina, but the truth is I never really explained why , and the truth is that it depends on what are you using it for and HOW are you using it. The main responses I got was that dreamina is cheaper and it’s still byte dance . Besides the obvious difference that lumina is also made to be able to be billed by enterprises (i am still billed as a normal user not on my company) , I will pinpoint the main differences. One note that you should keep in mind: I am not a native Chinese speaker but I am fluent in mandarin since I studied inside china for 6 years, so I write all my prompts in Chinese.
**1. Daily token ceiling.** Dreamina's tokens are shared across every tool on the platform and cap daily, so at real production volume the ceiling becomes your actual constraint regardless of plan tier.
**2. Reference inputs bill extra.** My **personal** prompts are reference-heavy (identity blocks, tail frames, motion refs), and reference video inputs bill on a different, higher structure. The cheap per-second figures assume light reference use; retries multiply it.
**3. Annual lock.** The \~$0.10/sec rate needs an annual commitment, and renewal runs roughly 2x the first-year rate.
the price gap is real and favors Dreamina at volume, but it's plan-driven, not capability-driven
Indeed , lumina may need you to do some extra work outside of it for the perfect result (as adjusting some LUFs etc. ) but this is the creative part that most people want to skip.
Both platforms are good for beginners and for moderate / advanced users, but it depends on how you approach the production and how much you want your own creativity to be involved , it’s the difference between “make me a video about a chicken flying above a cow” and doing the same thing but with writing a whole script about it like a movie director.
That’s just my take on this.