Product guide
HangarBench guide
Compare agents and models on your own tasks. In development.
In development
In development. HangarBench is not in the shipped app yet — there is nothing to set up today.
What it is
HangarBench will let you compare agents and models on your own tasks. You give it tasks, it runs different setups, and it reports scores so you can see which one does best on your real work.
What it will do
- Run the same task across different agents and models.
- Score each run so you can compare them fairly.
- See which setup gives you the best result on your own work.
Good to know
- Follow the roadmap and status pages to see when it goes live.