Same task, different models.

Browse every benchmark:

• Image: A person, and a business or shop.
• Video: A person, and someone at work.
• Rendered promo (MP4): Promotional videos written with HyperFrames or Remotion and rendered to MP4.
• Web page: One self-contained HTML page from a standard brief.
• Marketing: Flyers, carousels and social assets for small businesses.
• Logo: A logo in SVG, shown in its main versions.
• Web promo (HTML): A 6-second HTML, CSS and JS promo that loops.
• SVG drawing: A drawing made entirely from code.
• Simulation: Physics and motion that must behave correctly.
• Game: A complete playable game in one file.
• Workflows: Real workflows on saved inputs, by model and technique.
• Decision race: Every business rated P1–P4 in batches: speed, and agreement with the original research.
• Prospection challenge: Insight from the prospection database, the same questions each run.
• Deep research: Find out everything on a fixed subject.