I think Deno isn't included in the benchmark because it's a harder comparison to make than it might seem.

Deno's dependency architecture isn't built around npm; that compatibility layer is a retrofit on top of the core (which is evident in the source code, if you ever want to see). Deno's core architecture around dependency management uses a different, URL-based paradigm. It's not as fast, but... It's different. It also allows for improved security and cool features like the ability to easily host your own secure registry. You don't have to use npm or jsr. It's very cool, but different from what is being benchmarked here.

All the same, you can run deno install in a directory with a package.json file an it will resolve and install to node_modules. The process is also written in compiled code, like bun... so I was just curious.

edit: replied to my own post... looks like `deno install --allow-scripts` is about 1s slower than bun once deno.lock exists.