II.
Benchmark overview
Reference · livebenchmark:agentboard
AgentBoard overview
AgentBoard is an analytical benchmark and leaderboard for LLM agents covering embodied tasks, web, tool use, and games. Reports fine-grained progress and sub-goal metrics rather than only success rate.
Attributes
displayName
AgentBoard
homepageUrl
kind
agent-leaderboard
targetsKind
AgentVersion
description
AgentBoard is an analytical benchmark and leaderboard for LLM agents
covering embodied tasks, web, tool use, and games. Reports
fine-grained progress and sub-goal metrics rather than only success
rate.
Outgoing edges
applies_to1
- domain:ml-ai·DomainML/AI
Incoming edges
bounds_subject1
- scope-boundary:agentboard.scope·ScopeBoundary