The Open Agent Leaderboard

The Open Agent Leaderboard（开放智能体排行榜）

IBM Research published the Open Agent Leaderboard on Hugging Face to evaluate agents across language understanding, tool use, and multi-step reasoning tasks; the post does not disclose dataset size, model scores, or the evaluation date.

Why it matters: HKR-H and HKR-R pass because an open agent leaderboard speaks to agent-eval pain. HKR-K fails: the article lacks scores, dataset size, and evaluation date, so it sits at the featured threshold.