K-EXAONE

K-EXAONE (Reasoning) was released on December 31, 2025.

K-EXAONE (Reasoning) was created by LG AI Research.

K-EXAONE (Reasoning) scores 25 (estimated) on the Artificial Analysis Intelligence Index, placing it above average among other open weight models of similar size (median: 25).

Yes, K-EXAONE (Reasoning) is a reasoning model. It uses extended thinking or chain-of-thought reasoning to work through complex problems before providing an answer.

K-EXAONE (Reasoning) supports text input.

K-EXAONE (Reasoning) supports text output.

No, K-EXAONE (Reasoning) does not support image input. It can only process text.

No, K-EXAONE (Reasoning) is not multimodal. It only supports text input.

K-EXAONE (Reasoning) has a context window of 260k tokens. This determines how much text and conversation history the model can process in a single request.

Yes, K-EXAONE (Reasoning) is open weights. The model weights are publicly available and can be downloaded for self-hosting.

K-EXAONE (Reasoning) has 236 billion parameters (23 billion active).

K-EXAONE (Reasoning) is a Mixture of Experts (MoE) model with 236 billion total parameters, but only 23 billion active parameters are used during inference.

K-EXAONE (Reasoning) is released under the K Exaone license. This license allows commercial use.

K-EXAONE (Reasoning) achieves a score of 25 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.

K-EXAONE (Reasoning) has a knowledge cutoff of December 2024. The model's training data includes information up to this date.

K-EXAONE (Reasoning) is an open weights model that can be self-hosted.

K-EXAONE (Reasoning) is an open weights model that can be downloaded and self-hosted.