Beta version: *Information might not be fully accurate. Please report any discrepancies.

BigBench Extra Hard

bigbench-extra-hardIntelligenceHigher is better

Extra-hard subset of BIG-bench focusing on challenging reasoning and knowledge tasks.

Loading chart…

Loading leaderboard…

Benchmark Info

Current SOTA

Best Open Source

Median Score33.1%

DistributionBalanced

Score Distribution

P10

P50

P90

0100%

LiveBench

livebench

MMLU-Pro

mmlu-pro

HLE

hle

HLE-Full

hle-full

HLE-Full (w/ tools)

hle-full-tools

CritPt

critpt