/

← All benchmarks

BABILong

long-context · reasoning-in-haystack↗ Quelle

53Evidence

not yet measured

What it measures

Schliessen ueber verstreute Fakten in sehr langen Texten.

What it does not measure

Zusammenfassen, Schreiben.

How the score is composed

Currency50

Discrimination50

Headroom50

Contamination safety60

Gaming resistance55

Methodological rigor50

Reproducibility60

Details

Benchmaxxing risk: medium
Difficulty: high
Public test set: yes
Human-verified: no
LLM-judged: no
Skala: 0–100 % correct (höher besser)

Hand-maintained — no automatic adapter yet.

Best measured models

No results ingested yet.