AI Paper Atlas

Research map / Large language models / Reasoning and chain of thought

RADAR: Reasoning–Ability and Difficulty-Aware Routing in Language Models

Nigel Fernandez, Branislav Kveton, Ryan A. Rossi, Andrew Lan, Zichao Wang

ICLR 2026 · Large language models › Reasoning and chain of thought · approach: Search and self-correction in reasoning

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes RADAR, a difficulty-aware router that chooses both model size and reasoning budget for reasoning language models to trade off performance against cost.

Problem

Choosing a reasoning language model for deployment involves a performance-cost tradeoff across model size and reasoning budget, where larger models and larger budgets improve accuracy but increase cost and latency.

Approach

Routes each query jointly over model size and reasoning budget using reasoning-ability and difficulty awareness, rather than choosing a single model or fixed budget. Details beyond the abstract are not stated.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
route with RADAR instead of IRT-RouterIRT-Router · GPQA-Diamond in-distributionhypervolume: 0.6942 -> 0.7513Improved
route with RADAR instead of IRT-RouterIRT-Router · FRAMES out-of-distributionhypervolume: 0.8548 -> 0.8865Improved
route with RADAR instead of IRT-RouterIRT-Router · AIME out-of-distributionhypervolume: 0.7915 -> 0.7566Hurt
use RADAR to match 90% of o4-mini-high performanceo4-mini high budget · MATH-500cost fraction: 1.31% of costImproved
add Qwen3-14B configurations via adaptive testing (RADAR++)RADAR · FRAMES in-distributionhypervolume: 0.8762 -> 0.8830Improved
replace linear difficulty transform with two-layer MLP (ablation)RADAR · eight benchmarkshypervolume: GPQA-Diamond 0.7513 -> 0.7308; FRAMES 0.8762 -> 0.8777No clear effect
use Chebyshev instead of linear scalarization (ablation)RADAR · GPQA-Diamond out-of-distributionCPT(90%): 29.94% -> 17.60%Improved
use Chebyshev instead of linear scalarization (ablation)RADAR · in-distribution benchmarkshypervolumeNo clear effect
train on only 20% of subsampled training queries (ablation)RADAR · eight benchmarkshypervolumeNo clear effect
expose 5% of AIME and MATH queries in training (ablation)RADAR · AIME out-of-distributionhypervolume: 0.7566 -> 0.7665 (30%: 0.7787)Improved

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