Not tested by us yet. Wizshot is available only inside Wizstar's own web product, and we generate no video in our tests. Everything below comes from Wizstar's launch announcement on PR Newswire. We'll update this post when the model gets an API or independent comparisons appear.
Wizshot 1.0 is the first in-house video model from Wizstar, an AI content platform. It is a reference-to-video model: you give it up to nine reference images and a prompt, and it returns 720p clips of 5 to 15 seconds. Wizstar says quality is comparable to Wan 3.0 and MiniMax H3 at a generation cost 72% lower than Seedance 2.0. Those are the vendor's claims, with no benchmark published in the release.
| Wizshot 1.0 | |
|---|---|
| Maker | Wizstar |
| Announced | 2026-11-06 |
| Type | Reference-to-video |
| Inputs | Prompt plus up to 9 reference images |
| Resolution | 720p, landscape and portrait |
| Clip length | 5 to 15 seconds |
| Videos per generation | Up to 4 |
| Price | 5 credits per second (Credit Mode) |
| Access | Wizstar website, any logged-in user |
| API | None announced |
What is new
Reference-to-video means the model is built to keep things consistent across a clip: a product, a face, a character. Wizstar lists product consistency and human identity consistency among its optimization targets, alongside audio-visual synchronization, prompt adherence and physical plausibility. The announcement names marketing, e-commerce and social media as the intended uses.
The nine-image limit is the concrete feature. A shop owner can feed in front, side and detail shots of one product and ask for a short ad clip. Whether the model actually holds the product's shape and labels is exactly what the announcement does not show.
Pricing and access
Wizshot has two modes. Credit Mode costs 5 credits per second and gets higher queue priority. Unlimited Mode lets eligible annual and Plus Monthly subscribers (or above) generate without spending credits during a limited-time benefit period, subject to plan terms. Free accounts can use the model too.
The release gives no dollar price per credit, so we cannot convert 5 credits per second into a cost per clip, and we won't guess. The "72% lower than Seedance 2.0" claim is therefore also impossible to check from the announcement alone.
How it compares on paper
For a different video model with a published per-token price, Gemini Omni 1.2 Flash is listed on OpenRouter at $1.50 per million input tokens and $9 per million output tokens. That is a token price, not a price per second of video, so it does not map directly onto Wizshot's credits. Wan 3.0, MiniMax H3 and Seedance 2.0 are the models Wizstar names, and none is a like-for-like price comparison without Wizstar's credit rates.
What to check before you rely on it
Four things the announcement leaves open. First, whether the 5 to 15 second clips include generated audio, since it lists audio-visual synchronization as a goal but gives no format. Second, how identity consistency holds when a person appears in several reference images. Third, how often a generation needs a re-run, because that multiplies the real price per usable clip. Fourth, what the credit costs in dollars on each plan. Until clips and rate cards are public, treat the cost claim as marketing.
Who should care
- E-commerce and social teams who already use Wizstar and want product-consistent clips from existing photos. This is the audience the announcement targets.
- Developers. There is no API, so you can't script it yet.
- Anyone choosing a video model on quality. Wait for side-by-side clips. A vendor line saying "comparable" is not evidence.
Video prices are hard to compare because every vendor bills differently. Our post on LLM cost per feature shows how to turn a unit price into a per-task budget, and for another small release from the same week, see Underdog Woof, a 32 MB text-to-speech model.