Shared task
Benchmark brief
Create a one-minute explainer video that explains quantum superposition in simple terms for a general audience.
Test evidence · 2026-08-01
InVideo AI and InVideo Studio
No retained output
Two different products were tried
The August 1, 2026 InVideo attempt generated an approximately one-minute result in about 11 minutes, but the output was not retained for independent checking. The missing output, exact prompt, and project-usage record prevent a visual, cost, or overall rank.
Original result
The output was not preserved.
No output file or public share link was retained.
This result cannot be independently replayed or visually rescored.
Generation11 minutes
Output60 seconds
Normalized11 minutes / min
VisualNot scored
CostNot scored
Workflow
Two different products were tried
InVideo AI Agent is a conversational, scene-building workflow. InVideo Studio is a separate template and timeline editor with a separate subscription. Neither should be described as the other.
Review
Visual explanation finding
The run generated a result, but the output was not retained. Without the file or share link, no independent visual score can be assigned.
Review status: benchmark record unverified. See the visual-scoring rubric →
InVideo AI Agent is not a one-shot generator
InVideo's current help center describes Agent mode as a back-and-forth creative process and explicitly says it is not a one-shot generator. The same product also offers Autopilot for a single-shot flow. The attempted Agent result must therefore stay labeled as conversational and scene by scene.
InVideo Studio is a separate manual product
InVideo says Studio uses templates or a blank canvas with a timeline editor and requires a separate subscription from InVideo AI. The recorded attempt adapted a meditation-and-wellness template because no suitable free explainer path was available. That attempt was template editing, not an AI one-shot test.
Why generative InVideo projects can be expensive
InVideo provides access to 200+ media models, including Veo, Sora, Kling, and Seedance. It says generative models use provider list pricing, while Agent exchanges may invoke several image, video, voice, and reasoning operations. That makes total cost depend on the exact models and assets generated, not merely the final runtime.
What can and cannot be concluded
The run generated a result, but there is no retained output, so The Video Layer cannot defensibly claim it was better or worse than Synthesia. The visual score remains withheld.
Product fit
InVideo positions Agent mode around filmmaking, directors, brand films, cinematic scenes, and iterative creative control. It also offers an Explainer workflow, so it would be inaccurate to say the product cannot make explainers. The narrower finding is that this recorded attempt did not establish a useful quantum-superposition explainer.
Price record
Pricing basis
Current InVideo annual plans imply about $0.21–$0.22 per credit. InVideo publishes 80 credits/minute for Autopilot Pro and 160 credits/minute for Ultra, or roughly $16.67–$35.56 per minute if the annual allowance is fully used. This is reference arithmetic, not the tested Agent project's cost.
Open pricing screenshot captured 2026-08-01 →
Primary sources