A publicly available evaluation score is needed to verify whether the reproduction was successful.

[lycfight]

The models used in the paper are not publicly available. Could you provide the score produced by a publicly accessible model on agentless-mini, so that we can determine whether agentless-mini was correctly reproduced and executed?