Benchmark Almanac

K-EXAONE

K-EXAONE ranks first on WMT24++.

LG AI Research Released December 31st, 2025 Family K-EXAONE The lab's model card Compare with others
26
Benchmarks with a score
1
Ranked first, out of five or more models
1
In the top three
LG AI Research Dec 2025

K-EXAONE

Open weights Flagship

138.0 #194 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • IFBench 67.3% #51 /80
  • AA-LCR (Artificial Analysis Long Context Reasoning) 53.5% #29 /38
  • KMMLU-Pro 67.3% #6 /7

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
WMT24++ 90.5% #1 of 15 Lab model card
IFEval 89.7% #33 of 135 Lab model card
MMLU-Pro 83.8% #44 of 167 Lab model card
KoBALT 61.8% #1 of 3 Lab model card
MMMLU 85.7% #29 of 71 Lab model card
OpenAI MRCR 52.3% #6 of 10 Lab model card
IFBench 67.3% #51 of 80 Lab model card
Ko-LongBench (LG) 86.8% #2 of 3 Lab model card
OpenAI MRCR v2 52.3% #10 of 14 Lab model card
CLIcK 83.9% #3 of 4 Lab model card
AA-LCR (Artificial Analysis Long Context Reasoning) 53.5% #29 of 38 Lab model card
Humanity's Last Exam 13.6% #128 of 154 Lab model card
KMMLU-Pro 67.3% #6 of 7 Lab model card

Math and science

BenchmarkScoreRankSource
AIME 2025 92.8% #35 of 143 Lab model card
GPQA Diamond 79.1% #124 of 302 Lab model card
HRM8K 90.9% #2 of 3 Lab model card
HMMT Nov 2025 86.8% #19 of 28 Lab model card
IMO-AnswerBench 76.3% #32 of 37 Lab model card

Coding

BenchmarkScoreRankSource
LiveCodeBench v6 80.7% #29 of 90 Lab model card
SWE-bench Verified 49.4% #129 of 155 Lab model card
CodeUtilityBench (LG) 71.9% #1 of 1 Lab model card

Agents and tools

BenchmarkScoreRankSource
tau2-bench 73.2% #51 of 79 Lab model card
BrowseComp 31.4% #75 of 81 Lab model card
Terminal-Bench 2.0 29% #60 of 63 Lab model card

Safety and security

BenchmarkScoreRankSource
KGC-Safety (LG) 96.1% #2 of 2 Lab model card
WildJailbreak 89.9% #2 of 2 Lab model card

Its lab's numbers put it at 138.0, number 194 of 404

The Self-Reported Index fits one score to the 18 capability results LG AI Research reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 136.1 to 139.9. How we compute it.

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

K-EXAONE-236B-A23B model card (Hugging Face) Dec 31, 2025

Its system card

Every result in the longer documents the lab published with it. 2 of them aren't in the launch post.

K-EXAONE Technical Report Jan 5, 2026

Capabilities

Safety