II.
TestSet overview
Reference · livetest-set:gaia-validation
GAIA validation split overview
Validation split of the GAIA benchmark (Mialon et al., 2023). 165 held-out questions across three difficulty levels. Used as the public-leaderboard split because the test split is hidden.
Attributes
displayName
GAIA validation split
benchmarkId
splitName
validation
itemCount
165
description
Validation split of the GAIA benchmark (Mialon et al., 2023). 165
held-out questions across three difficulty levels. Used as the
public-leaderboard split because the test split is hidden.
sourceUrl
Outgoing edges
split_of1
- benchmark:gaia·BenchmarkGAIA
Incoming edges
uses_test_set1
- eval-run:gaia.claude-code.2025·EvalRun