AI Paper Atlas

Research map / Large language models / AI and society

When AI Benchmarks Plateau: A Systematic Study of Benchmark Saturation

Mubashara Akhtar, Anka Reuel, Prajna Soni, Sanchit Ahuja, Pawan Sasanka Ammanamanchi, Ruchit Rawal, Vilém Zouhar, Srishti Yadav, Chenxi Whitehouse, Dayeon Ki, Jennifer Mickel, Leshem Choshen and 25 more

ICML 2026 · Large language models › AI and society · approach: Social-science view of AI evaluation · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase benchmark age (past 24 months to over 60 months)LLM benchmarks · 60 text-based LLM benchmarksproportion saturated: 42.9% -> 54.5%Hurt
use a private held-out test set instead of public test dataLLM benchmarks · 60 benchmarks (56 public, 4 private)S_indexNo clear effect
use open-ended instead of closed-ended output formatLLM benchmarks · 60 benchmarks (28 closed, 31 open)S_index: age-balanced, p=0.40No clear effect
use multilingual instead of English-only benchmarksLLM benchmarks · 60 benchmarks (16 multilingual, 44 English-only)saturation rate: mean age 32.9 vs 48.9 monthsNo clear effect
use non-templated instead of templated benchmarksLLM benchmarks · 60 benchmarks (46 non-templated, 14 templated)saturation behaviour: p=0.10No clear effect
use expert-curated instead of crowdsourced benchmarksLLM benchmarks · 60 benchmarkssaturationImproved
larger test setsLLM benchmarks · 60 benchmarks, joint regressionS_indexImproved
higher citation counts, controlling for ageLLM benchmarks · 60 benchmarksS_index: rho = 0.22, p = 0.12No clear effect
smaller top-k (k=3) in saturation index (ablation)Saturation index · sensitivity analysisvariance of estimateHurt
set alpha=1 in effective test size (ablation)Saturation index · sensitivity analysisdependence on test set sizeHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy