Benchmark Almanac

K-EXAONE 2.0

K-EXAONE 2.0 ranks first on 2 of the 20 benchmarks with five or more models scored.

LG AI Research Released July 31st, 2026 Family K-EXAONE The lab's model card Compare with others
26
Benchmarks with a score
2
Ranked first, out of five or more models
4
In the top three
LG AI Research Jul 2026

K-EXAONE 2.0

Open weights Flagship

145.0 #131 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • IFBench 72.6% #31 /80
  • SciCode 40.1% #29 /49
  • AA-LCR (Artificial Analysis Long Context Reasoning) 56.2% #28 /38

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
OpenAI MRCR 94.4% #1 of 10 Lab model card
IFEval 92.4% #18 of 135 Lab model card
OpenAI MRCR v2 94.4% #2 of 14 Lab model card
MMLU-Pro 83.5% #48 of 167 Lab model card
MMMLU 86.6% #22 of 71 Lab model card
Ko-LongBench (LG) 89.6% #1 of 3 Lab model card
IFBench 72.6% #31 of 80 Lab model card
KMMLU-Pro 69.1% #3 of 7 Lab model card
CLIcK 84.2% #2 of 4 Lab model card
Humanity's Last Exam 18.3% #110 of 154 Lab model card
AA-LCR (Artificial Analysis Long Context Reasoning) 56.2% #28 of 38 Lab model card
Global MMLU Lite 3.0 86.6% #4 of 4 Lab model card

Math and science

BenchmarkScoreRankSource
PolyMath 71.3% #5 of 24 Lab model card
GPQA Diamond 82.2% #104 of 302 Lab model card
HRM8K-KSM 91.1% #1 of 2 Lab model card
AIME 2026 92.3% #24 of 38 Lab model card
IMO-AnswerBench 78.6% #28 of 37 Lab model card
HMMT 2026 78.4% #18 of 19 Lab model card

Coding

BenchmarkScoreRankSource
SWE-bench Verified 68.2% #90 of 155 Lab model card
SciCode 40.1% #29 of 49 Lab model card

Agents and tools

BenchmarkScoreRankSource
Claw-Eval 80% #1 of 13 Lab model card
GAIA 23.3% #52 of 64 Official leaderboard
τ³-bench Banking 14.2% #19 of 22 Lab model card
Terminal-Bench 2.1 43.8% #52 of 56 Lab model card

Safety and security

BenchmarkScoreRankSource
KGC-Safety (LG) 99.8% #1 of 2 Lab model card
ROK-FORTRESS 89.5% #1 of 1 Lab model card

Its lab's numbers put it at 145.0, number 131 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 141.4 to 148.6. How we compute it.

Its launch post

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

K-EXAONE-2.0-750B-A37B model card (Hugging Face) Jul 31, 2026

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 2.0 Technical Report Aug 5, 2026

Capabilities

Safety