readable AI benchmarks (simplified)

readable AI benchmarks

LG AI Research

K-EXAONE 2.0

reasoning · open weights · Aug 12, 2026

K-EXAONE 2.0 has a Quality Score of 25.9, ranking 205th among 283 scored models. It has a 262,144 token context window. Its Reliability Score is 44.3, ranking 179th among 288 scored models, below average (total average: 48.2).

Quality Score25.9coverage 100.0%: AA Intelligence and AA-Omniscience outcomes available
Value Scorenot available
Factual reliability44.3
Cache Discountnot available

Model specification

Reasoning
reasoning
Input modalities
none
Output modalities
none
Context window
262,144 tokens
Weights
open weights

Token prices USD per 1M tokens

Input
not available
Output
not available
Cached Input
not available
Cache write
not available
Cost per task
not available

Capability

Intelligence
7.6
Omniscience (native net)
-11.5
Correct
14.5%
Blended price
not available

Answer outcomes

Fully graded outcomes (Correct + Incorrect): 40.5%

Correct 14.5%Incorrect 26.0%Partial / not attempted 59.5%

Artificial Analysis benchmarks aa-v5

Current Intelligence Index components

AA-Briefcase v1.1
651.9 Elo
AutomationBench-AA
3.9%
private 657-task split from dataset v1.0.6
GDP.pdf
4.2%
FrontierCode v2.0
0.8%
AA Stirrup · 132 tasks, 3 attempts per task
Terminal-Bench v4.0
0.0%
AA mini-swe-agent standardized runs · 66 tasks, pass@1 averaged over 3 repeats per task
AA-Omniscience v1.1
-11.5 native net
HLE-Diamond
5.7%
CritPt v2 Hard
4.0%
AA Stirrup · hard subset, 40 of 70 problems
Terminal-Bench-Science v0.1.0
0.0%

Standalone evaluations

GDPval-AA v2.1
815.1 Elo
GPQA Diamond
82.9%
τ³-Banking
11.5%

Historical evaluations

SciCode v1.0.1
42.0%
AA Long Context Reasoning v1.1
60.0%
Humanity's Last Exam
18.6%
CritPt
0.9%
Terminal-Bench v2.1
40.4%

Similar models

Support me! Patreon