Benchmarks of K EXAONE15
BenchmarkByScoreRank
Overall1
Knowledge4
Science2
Coding1
Agents1
Documents1
Legacy · saturated or no longer reported5
No benchmark matches
—per task?
self-host onlyprice
—output speed?
262Kcontext
Rankingin thinking mode?
Alternatives?
ModelScoreCostPer taskWhat changes
ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
ModelScoreCostPer taskWhat changes
Run it yourself
estimateFits on✗ 8 GB (phone, laptop)✗ 16 GB✗ 24 GB GPU✗ 48 GB GPU✗ 80 GB GPU
Licencek-exaonecommercial use with conditions
History
2 versionsBenchmarks
all categoriesChat & Knowledge
Humanity's Last Exam13.9#149
AA Omniscience Index-58#246
CritPt (physics)1.1#129
AA Intelligence Index 4.3.2≈14.4#151
AA Omniscience · Accuracy16.4#214
AA-LCR (long context) 1.161.3#153
GPQA Diamond78.3#124
AIME 2025 · AA archived90.3#15
Coding
Terminal-Bench 2.130.3#96
LiveCodeBench · AA archived76.8#26
Agents & Tool Use
τ²-bench · AA74.3#90
Writing & Language
IFBench (instructions)64.7#72
FAQ
How much does K EXAONE cost?›
There is no per-unit API price: self-host only.
Is K EXAONE free to use?›
There is no free API. The weights are open, so you can run it yourself.
How good is K EXAONE, and what is it best at?›
#148 of 334 in Chat & Knowledge, #84 of 239 in Coding, #122 of 175 in Agents & Tool Use, #72 of 245 in Writing & Language.
Is K EXAONE better than DeepSeek V3.2 Speciale?›
On AA Intelligence Index 4.3.2, K EXAONE scores 14.4 vs 14.5 — about the same.
What is the context window of K EXAONE?›
262K tokens.
When was K EXAONE released? Is there a newer version?›
Released 2025-12-31. It has been superseded by K-EXAONE 2.0 0803.
Is K EXAONE open source?›
Yes, the weights are open (k-exaone licence).
Can I run K EXAONE locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can K EXAONE read images?›
No, it accepts text only.
What is K EXAONE?›
K EXAONE is an open-weights language model from LG AI Research, released Dec 2025.
What are the best alternatives to K EXAONE?›
GPT-6 Luna (better), GLM 5.3 Flash (open weights).