EXAONE 2.0
EXAONE 2.0
Korea's largest published model, and a serious piece of national AI infrastructure.
Updated
The read
LG released EXAONE 2.0 at a scale no Korean lab had published before, triple the size of its predecessor, with the weights on Hugging Face. It is strongest in Korean, where it is the reference open model, and respectable in English. It is brand new, the tooling around it is immature, and running something this size is a datacentre decision rather than a workstation one.
Where it shines
- The reference open model for Korean
- Published openly despite its scale
- Backed by a company that will keep maintaining it
Worth knowing first
- Very new, so the surrounding tooling is immature
- Serving it needs a datacentre, not a workstation
Index score
77.8
Provisional. This score is read from the product's public capability, not from the one-prompt rebuild the fully reviewed entries went through, so treat it as a placing rather than a verdict.
- Rank
- 32 of 53
- Builds
- LLMs
- Category
- Open weights
- Output
- Published weights at national scale
- Pricing
- Free tier
- Best for
- Korean-language work on weights you can hold
Scores are our own editorial judgement, weighted craft 55, speed 10, control 15 and value 20.
Also worth a look
Full index- 05Kimi K3Kimi K3ProvisionalThe largest open-weights model anyone has shipped, and it reads a million tokens without blinking.Open weightsFree tier87.5
- 06GLM-5.2GLM-5.2ProvisionalFrontier-class output, an MIT licence, and a million tokens of context. That combination is the story.Open weightsFree tier87.1
- 12Qwen 3.6Qwen 3.6ProvisionalThe open Qwen release most people should actually run, in sizes that fit real hardware.Open weightsFree tier83.3
