Run Frontier-Bench

Install Harbor

Frontier-Bench runs in Harbor.

uv tool install 'harbor[modal]'

Run Frontier-Bench

Frontier-Bench contains 4 tasks that require GPUs so you need to run it with a sandbox that has GPU access. We used Modal for our experiments. Daytona also recently added GPU support and we hope others will follow suite.

harbor run -d frontier-bench/frontier-bench \
  -e modal \
  -a claude-code \
  -m anthropic/claude-sonnet-5 \
  -n 32

Analyze results

You can run the Harbor viewer if you want to inspect your experiment results.

harbor view jobs

Upload results

You can also upload your results to Harbor Hub to store for later or share with others. All uploads are private by default.

harbor upload jobs/<name> # or harbor run ... --upload

Join our community

If you enjoy the benchmark or have feedback for how we can make it better, join our Discord and let us know!