K-EXAONE Technical Report

Benchmark Model Rank Results
agents-on-bench-telecomK-EXAONE (236B, reasoning)#7Accuracy: 73.5
mathematical-reasoning-on-aime-2025K-EXAONE (236B, reasoning)#15Accuracy: 92.8
code-generation-on-livecodebench-v6K-EXAONE (236B, reasoning)#16Pass@1: 80.7
instruction-following-on-ifbenchK-EXAONE (236B, reasoning)#23Accuracy: 67.3
multi-task-language-understanding-on-mmlu-proK-EXAONE (236B, reasoning)#26Accuracy: 83.8
deep-research-agents-on-browsecompK-EXAONE (236B, reasoning)#51Accuracy: 31.4
multiple-choice-question-answering-mcqa-on-gpqa-diamondK-EXAONE (236B, reasoning)#57Accuracy: 79.1
code-generation-on-swe-bench-verifiedK-EXAONE (236B, reasoning) (mini-SWE-agent)#75Accuracy: 49.4