OpenAI's Sora 2 is the flagship in text-to-video. It generates 5- and 10-second high-res videos with realistic physics and stable camera work. It's available in Quantium without a separate subscription.
In just 8 minutes, we'll break down how to write a Sora prompt to get your video right on the first or second try. You'll find a working template at the end.
1Open the Video Section
In the bot: «🎬 Video» → «Sora 2». One 5-second video costs roughly 15 credits for Standard and 25 for Pro quality. Standard's fine for a test run.
2Video Prompt Formula
Sora works best when your prompt describes: (1) what's in the shot — subject and setting, (2) what's happening — action or movement, (3) how it's shot — angle and camera movement, (4) the atmosphere — lighting, weather, time of day.
3Describe Camera Movement
This is critical — without clear instructions, the camera might just 'freeze'. Use: slow zoom in, tracking shot from the left, handheld camera following, aerial drone shot rotating around, static wide shot.
4Choose Duration and Format
5 seconds — for short, impactful shots (b-roll, reaction shots). 10 seconds — for storytelling or complex action. Format: 16:9 for YouTube/web, 9:16 for Reels/Shorts/TikTok, 1:1 for Instagram.
5Refine Your Prompt Based on Results
If your first video's subject isn't facing the right way, add facing the camera. If the movement's too jerky, try slow steady motion. If the shot's too dark, use bright daylight, well lit. Sora's great at picking up on refinements — usually 2-3 iterations get you that 'just right' video.
What's Next
Sora 2 is powerful, but it's pricey. Don't just render 'on a whim' — think through your prompt first. Here's a pro tip: ask ChatGPT in Quantium to rewrite your text draft into Sora's format — it often saves you 1-2 iterations.
Ready to try it out? 20 free credits are enough for 3-5 generations.
Open @quantium_aibot →
