llmbench

What problem does it solve?

This skill eliminates the guesswork in comparing large language models by automating a full benchmark pipeline that runs a given prompt across four models in parallel and produces a shared, visual comparison video.

Core Features & Use Cases

• Parallel model execution across four model endpoints using OpenCode to speed up benchmarking.
• Remotion-based 2x2 video grid that visually juxtaposes outputs for easy assessment.
• Use cases include prompt design evaluation, model capability comparison, and end-to-end workflow testing across model families.

Quick Start

Run the llmbench pipeline with a prompt to test four models in parallel and render the comparison video.