Luma launches Uni-1.1 API for teams

Hot take: this is less about a flashy new image model and more about Luma productizing controllable creative workflows for teams that care about consistency, not just one-off generations.

• –The strongest signal is the API shape: reference-heavy generation, natural-language editing, and SDK support make this feel built for pipelines, not hobbyist prompting.
• –Luma is clearly leaning into "reason before render" as its differentiator; that claim matters most if it materially improves character consistency, layout fidelity, and brand control.
• –The reported ~31 second generation time means this is probably best for high-value creative work where quality beats speed, not instant consumer sketching.
• –The marketing language is strong, but the real test is whether the model holds up under messy, multi-step production briefs better than stitched-together image stacks.