Inference
EN
LG AI Research

K-Exaone

K-Exaone is a reasoning model from LG AI Research. 9 benchmarks count toward its index, in 5 categories.

  • Closed weights
  • Reasoning
  • Input: text
  • Released 2025-12
RankedEnough results for an index.
IndexAll categories, out of 100.42.3±7.9Median 51.0
CoverageIndex weight that has results.75%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.256KMedian 262K
EloLMArena rating and rank.—Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.2 counted43.1167 / 218
CodingCode writing and repair.1 counted37.7—
ReasoningLogic problems and puzzles.2 counted44.2145 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.3 counted39.6196 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted49.5—
MathMath problems.None counted——

Results

9 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Artificial Analysis IFBenchInstruction64.7%56 / 146
τ²-Bench Tool-Agent-User EvaluationAgentic74.3%76 / 149
Artificial Analysis GPQA DiamondKnowledge78.3%104 / 199
Critical Physics TasksReasoning1.1%118 / 214
Artificial Analysis Humanity's Last ExamKnowledge13.9%129 / 222
Artificial Analysis Long Context ReasoningReasoning61.3%140 / 217
GDPval-AA normalizedAgentic0.0%101 / 133
Software Engineering Benchmark VerifiedCoding49.4%96 / 114
Artificial Analysis Omniscience AccuracyKnowledge16.4%146 / 164
Artificial Analysis AIME 2025Math90.3%—

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index43.0 → 42.3
23 Oct 202612 Nov 2026

Questions

How good is K-Exaone?
Its index is 42.3 out of 100. That is place 196 of 283 ranked models.
What is K-Exaone best at?
Reasoning: 44.2 out of 100, place 145 of 227.
What is the context window of K-Exaone?
256K tokens.
Is K-Exaone open source?
No. Its weights are not published.
When was K-Exaone released?
December 2025.

Badge

K-Exaone on Inference360: place 196 of 283, index 42.3
HTML
<a href="https://inference360.ai/models/k-exaone"><img src="https://inference360.ai/badges/k-exaone.svg" alt="K-Exaone on Inference360: place 196 of 283, index 42.3" height="20"></a>
Markdown
[![K-Exaone on Inference360: place 196 of 283, index 42.3](https://inference360.ai/badges/k-exaone.svg)](https://inference360.ai/models/k-exaone)

Compare with

Trinity-Large-Preview42.2DeepSeek V3.1 (Reasoning)43.0K-EXAONE 2.048.8Exaone 4.0 32B32.4

More from LG AI Research

All models
Exaone 4.0 32B32.4Exaone 4.0 1.2BUnrankedK-EXAONE 2.048.8