AI Paper Atlas

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

BeyondBench: Benchmark-Free Evaluation of Reasoning in Language Models

Gaurav Srivastava, Aafiya Shamshad Hussain, Zhenyu Bi, Swastik Roy, Priya Pitre, Meng Lu, Morteza Ziyadi, Xuan Wang

ICLR 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks

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
add explicit algorithm description (ablation)GPT-OSS-120B · Tower of Hanoi (n=7)accuracy: 18.4% -> 24.2%Improved
add explicit algorithm description (ablation)GPT-OSS-120B · Sudoku (9x9)accuracy: 8.2% -> 12.1%Improved
add explicit algorithm description (ablation)GPT-OSS-120B · N-Queens (n=8)accuracy: 22.3% -> 28.7%Improved
add explicit algorithm description (ablation)GPT-OSS-120B · Boolean SAT (5 vars)accuracy: 38.1% -> 42.8%Improved
add tools (all tools combined) (ablation)GPT-5 · BeyondBench overallaccuracy drop when tools removed: 16.81% drop without toolsImproved
add tools (all tools combined) (ablation)GPT-5-nano · BeyondBench overallaccuracy drop when tools removed: 43.95% drop without toolsImproved
add thinking variantQwen3-4B · BeyondBench overallaccuracy: 58.63% -> 60.77%Improved
add thinking variantQwen3-30B-MOE · BeyondBench overallaccuracy: 69.00% -> 69.20%No clear effect
add mathematical fine-tuningQwen2.5-72B · BeyondBench overallaccuracy: 53.36% -> 48.48%Hurt
add mathematical fine-tuningQwen2.5-7B · BeyondBench overallaccuracy: 36.12% -> 31.63%Hurt
scale Qwen3 from 0.6B to 1.7B parametersQwen3 · BeyondBench overallaccuracy: 28.21% -> 45.75%Improved
scale Qwen3 from 14B to 32B parametersQwen3 · BeyondBench overallaccuracy: +1.79 pointsNo 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