The decision benchmark accepts any server that speaks the System One wire format, with no code: thinkless bench decisions run --provider systemone:<url>. Self-hosted System One models such as Kev or OpenJev are not on the leaderboard yet.
What to do
Run one of them locally or on a server you control, run the benchmark against it, verify the result, and open a pull request with the file under benchmarks/decisions/results/submitted/ and the regenerated leaderboard page. Say which model and version you ran and on what hardware.
The submission guide has every step. Questions are welcome in the Benchmark submissions discussions.
The decision benchmark accepts any server that speaks the System One wire format, with no code:
thinkless bench decisions run --provider systemone:<url>. Self-hosted System One models such as Kev or OpenJev are not on the leaderboard yet.What to do
Run one of them locally or on a server you control, run the benchmark against it, verify the result, and open a pull request with the file under
benchmarks/decisions/results/submitted/and the regenerated leaderboard page. Say which model and version you ran and on what hardware.The submission guide has every step. Questions are welcome in the Benchmark submissions discussions.