AI Paper Atlas

Research map / Large language models / Agents, code and math

RankLLM: Weighted Ranking of LLMs by Quantifying Question Difficulty

Xingjian Hu, Ziqian Zhang, Yue Huang, Kai Zhang, Ruoxi Chen, Yixin Liu, Qingsong Wen, Kaidi Xu, Xiangliang Zhang, Neil Zhenqiang Gong, Lichao Sun

ICLR 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 1 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
use EIP difficulty instead of simple error-count rank for question difficultyquestion difficulty estimation · Human difficulty alignmentconsensus agreement (%): 62.9 -> 90.0Improved
use EIP instead of 1PL/2PL/Multi-IRTquestion difficulty estimation · Human difficulty alignmentconsensus agreement (%): 52.9 -> 90.0 (vs Multi-IRT)Improved
use a mixed-size model pool instead of a single-size pool (ablation)EIP · Human difficulty alignmentconsensus agreement (%): 38.6%-64.3% -> 90.0%Improved
use a mixed-size model pool instead of homogeneous pools (ablation)EIP · difficulty distributiondead nodes: reduced by 83%; extreme cases below 3%Improved
replace IRT with EIPdifficulty/competency estimation · Convergence timetime (s): 1782.75 -> 0.00597 (vs 1PL IRT)Improved
remove half of the model pool (15 of 30) (ablation)EIP · Question difficulty rank stabilitySpearman rho: 0.9978 (k=1) -> 0.9382 (k=15)Mixed
drop one dataset (HellaSwag or MMLU-Pro) from the pool (ablation)EIP model ranking · Leave-one-dataset-outSpearman rho: 0.9604 / 0.9626 (vs 0.9973 for small datasets)No clear effect
weight correct answers by question difficulty (EIP) instead of plain accuracy/IRT in simulationmodel ranking · Simulated 5 models, 100 questionsagreement with ground-truth ranking: only EIP matched both orderingsImproved
use open-weight model pool only for difficulty estimation (ablation)EIP · Difficulty correlation with full poolSpearman: 0.96 (open-weight) vs 0.78 (proprietary)No clear effect

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