LG AI Research releases the open-weight K-EXAONE 2.0
LG AI Research publishes the K-EXAONE 2.0-750B-A37B model weights and technical report.
What happened
Event details
K-EXAONE 2.0 is a multilingual text foundation model using an MoE architecture, with 750B total parameters, about 37B active parameters per token and up to 256K tokens of context. Its official model card includes local and server deployment examples for text-message input and text generation.
Assessment
Why it matters
With 750B total parameters, about 37B active parameters and 256K context under Apache 2.0, the model enters the large-scale open-weight language-model tier and supports deployment and derivative work. Limited official hosted inference at launch makes its market and durability scores more cautious.
Availability
Access notes
The weights are public under Apache License 2.0. The inference-provider status shown on the model card does not affect public access to the weights.