A giant humanoid

A giant humanoid, made of fluffy blue cotton candy, stomping on the ground, and roaring to the sky, clear blue sky behind them.

Why this prompt works

This was generated on Runway Gen-3 Alpha. The prompt reads like a shot brief rather than a keyword list — it front-loads the subject, then layers in setting, motion, and lighting so the model has an unambiguous scene to render. Notice the explicit lighting and camera language. On Runway Gen-3, naming the light source and mood ("warm glowing neon", "dramatic lighting") gives the diffusion model a strong prior and is one of the biggest levers on perceived quality. Character and anatomy prompts are the highest-risk category on every model. The specificity here — wardrobe, posture, framing — narrows the space the model has to hallucinate extra limbs or warped faces. The honest takeaway: prompt specificity is doing most of the work. There's no magic token — it's a clear subject, a defined camera and light setup, and one concrete action, which is exactly what Runway Gen-3 needs to stay coherent.