Generation Models
Every video you generate in VidRush is shaped by two choices: the visual model that determines what your viewers see, and the reasoning effort that determines how deeply the AI researches and scripts your content. Together, they control both the quality and cost of each video.
This page walks you through each option, what it costs, and a practical strategy for scaling without wasting credits.
Visual Models
The visual model controls the type of media in your video: whether it uses images only or a mix of video and images.
Deep Video v1 Mini (Image)
Mini generates videos using high-quality images sourced from the web, combined with motion graphics. It does not include video B-roll.
When to use it: Testing new niches, topics where images work naturally (horror, history, storytelling, explainers), or when you want to keep costs low while validating an idea. Some audiences actually prefer an image-driven style. Don’t assume video is always better.
Deep Video v1 Pro (Video)
Pro is the full visual experience. It blends premium stock video B-roll with images and motion graphics, giving your videos more movement and visual variety.
When to use it: Main channel content where movement matters, high-retention videos, and topics where viewers expect to see real footage: travel, sports, documentaries.
Reasoning Effort
Reasoning effort determines how much processing power the AI dedicates to researching your topic, writing the script, and structuring the narrative. It applies to whichever visual model you choose.
Low Reasoning
Uses a lighter AI model optimized for speed and cost-efficiency. Research depth is lower, so it works well for general topics but may fall short on subjects that require in-depth analysis or academic-level accuracy.
Medium Reasoning
Uses VidRush’s most capable AI models. Research is deeper, scripts are more nuanced, and the output is noticeably stronger on complex topics. When paired with Pro, it maximizes video B-roll usage for the richest visual experience.
High Reasoning, Coming Soon. A higher-effort tier is in development for creators who need maximum research depth and script quality.
Credit Costs
Your final cost depends on the combination you choose. Credits are charged per minute of generated video.
| Visual Model | Reasoning Mode | Cost (Credits/Min) | Notes |
|---|---|---|---|
| Mini (Image) | Low | 32 | Lowest cost. Best for rapid testing. |
| Mini (Image) | Medium | 40 | High-quality image storytelling. |
| Pro (Video) | Low | 44 | Video B-roll included, but used more sparingly. |
| Pro (Video) | Medium | 55 | The Standard. Maximum quality and video density. |
🚀 The “Bootstrap” Scaling Strategy
You don’t need to start with the most expensive settings. The most successful VidRush creators follow a deliberate scaling path: test cheap, then upgrade once results prove a topic works.
Phase 1: Validate
Phase 1: Validate. Start with Mini + Low Reasoning (32 credits/min)
This is the fastest and cheapest way to find out if a topic gets clicks and holds attention. At roughly 1.00 per minute of video, you can test multiple ideas without draining your budget.
The goal here is discovery, not perfection. If the video gets views and engagement, you know the topic has potential.
Creator Tip: “If your niche doesn’t strictly need video clips, Mini + Medium often gives better outputs than Pro + Low. I only recommend Pro + Medium for creators who are already making money with VidRush. That’s when the investment pays for itself.”