Intelligence provides a platform that lets users generate and transform code, images, video, audio, slides, SVG, ASCII and other media using AI models. It features leaderboards across domains such as game development, mobile development, and web design, allowing AI labs to showcase and benchmark their models. The service supports both agentic and non‑agentic workflows for creating UI components, data visualizations, 3D designs, and more.
Funding
Funding not disclosed
Founders
Product
Problem
AI developers and research labs lack a unified, realistic benchmark to evaluate how well large models handle diverse multimodal tasks such as code generation, design, video, audio, and 3D rendering. Without comparable metrics, it is hard to gauge progress, select models, or demonstrate improvements to stakeholders.
Solution
Intelligence provides a cloud‑based evaluation platform that runs standardized, high‑fidelity tests across dozens of content types—including code, images, video, audio, SVG, ASCII, and 3D scenes. Models are scored on leaderboards that rank performance by task category and overall capability, giving labs an objective reference point. The platform supports both agentic (model‑driven) and non‑agentic (prompt‑only) workflows, allowing users to benchmark full‑stack pipelines or isolated components. Results are publicly visible and can be integrated via APIs, enabling continuous monitoring, model selection, and public validation of new releases. Intelligence also offers custom evaluation services for labs that need offline or proprietary test suites.
Target Audience
Primary customers are AI research labs, model developers, and enterprises that need rigorous, comparable performance metrics for multimodal generative models.
Features
- Multimodal benchmark suite covering code, design, video, audio, SVG, ASCII, and 3D assets
- Separate leaderboards for agentic and non‑agentic model configurations across each modality
- Automated evaluation pipeline that generates reproducible scores and rankings
- Public API for retrieving leaderboard data and submitting custom model evaluations
- Support for custom offline evals and bespoke test creation for enterprise labs
- Integration with major AI labs (e.g., OpenAI, Google DeepMind, xAI) for model release announcements