BenchmarkMD — AI Agent Benchmarking Publication
An editorial site for independent AI-agent benchmarks and cost analysis, built as an MDX publishing platform.

Real screenshot of the public page of benchmarkmd.ai, taken on a phone by the DevForge owner on 12 October 2026.
At a Glance
Tech Stack
The Challenge
The project's stated aim is independent testing of AI agents in real-world scenarios, with the API cost analysis that vendor benchmarks leave out. That calls for a publication where each report is easy to write, version and read.
What Was Built
A Next.js 16 App Router site where reports are MDX files with front matter, parsed with gray-matter and rendered with next-mdx-remote and GitHub-flavoured Markdown, styled with Tailwind CSS and deployed on Vercel with Vercel analytics and speed insights.
Where It Stands
Public repository. The reports' own statistics are the publication's editorial claims and are not repeated here.
Have a project like this in mind?
Let's discuss how we can engineer a solution for your specific business challenges.
Start Your Project