Benchmark family

WebArena

A family groups the versions of one benchmark. It is not evaluated and has no applicable models of its own; each of its 7 definitions is resolved separately.

AgenticLong HorizonMultimodalWeb AgentWeb browser

Definitions

Compare the newest 4 →

Newest first, following the recorded successor chain. Each definition is resolved against models separately.

BenchmarkST-WebAgentBenchICLR 2026
ActiveNewest in family
Derived from WebArena
Scale
222 counts
Primary metric
Completion Under Policy
Applicable models
51
BenchmarkVisualWebArena2024
Active
Derived from WebArena
Scale
3 counts
Primary metric
Task Success Rate
Applicable models
51
BenchmarkWebArenanamed benchmark
Active
Scale
241 counts
Primary metric
Task Success Rate
Applicable models
51
BenchmarkWebArena-Pro2026
Active
Related to WebArena
Scale
20 applications
Primary metric
Task Success Rate
Applicable models
51
BenchmarkWebArena-Verifiedverified
Active
Derived from WebArena
Scale
812 tasks
Primary metric
Task Success Rate
Applicable models
51
BenchmarkWebArena-Verified Hardcurrent hard subset
Active
Subset of WebArena-Verified
Scale
137 tasks
Primary metric
Task Success Rate
Applicable models
51
BenchmarkWebChoreArena2025
Active
Derived from WebArena
Scale
532 tasks
Primary metric
Task Success Rate
Applicable models
51

Family methodology

No family-level methodology recorded.

Across definitions

Tasks: Policy Constrained Web Interaction, Web Navigation, Web State Modification, Visually Grounded Web Interaction, Web Information Retrieval, Long Horizon Computer Use
Sources: None recorded at family level
Last reviewed: 2026-09-23

Applicability does not mean evaluated; benchmark result/model score data is outside this UI contract.