LG AI Research has expanded its footprint in the artificial intelligence sector by unveiling EXAONE Deep, a sophisticated reasoning model designed to tackle complex problems in mathematics, science, and coding. This latest development underscores LG’s commitment to advancing AI capabilities and positioning itself among leading technology innovators.
Previous iterations of LG’s AI models have shown promise, but EXAONE Deep marks a significant enhancement in performance and efficiency. By integrating advanced algorithms and optimizing model architecture, LG AI aims to set new benchmarks in the competitive AI landscape.
Maths Prowess
EXAONE Deep demonstrates exceptional mathematical reasoning across its 32B, 7.8B, and 2.4B model sizes. The 32B model achieved a score of 94.5 in general mathematics competency tests and matched the performance of the larger DeepSeek-R1 model in the AIME 2025 exam, highlighting its efficient learning and logic capabilities.
Science and Coding Excellence
In the realms of science and coding, EXAONE Deep continues to lead. The 32B model scored 66.1 on the GPQA Diamond test for doctoral-level science problem-solving and 59.5 on the LiveCodeBench coding proficiency evaluation. Similarly, the smaller models secured top positions in their respective benchmark categories, demonstrating versatility across different domains.
Enhanced General Knowledge
Beyond specialized reasoning, EXAONE Deep exhibits robust general knowledge understanding. The 32B model achieved a score of 83.0 on the MMLU benchmark, making it the best-performing domestic Korean model in this comprehensive assessment. This broadens the model’s applicability across various subjects and enhances its overall utility.
LG AI Research believes that EXAONE Deep’s reasoning advancements represent a leap towards a future where AI can tackle increasingly complex problems and contribute to enriching and simplifying human lives through continuous research and innovation.
As EXAONE Deep gains international recognition, including its inclusion in Epoch AI’s ‘Notable AI Models’ list, LG AI Research solidifies its position as a key player in the global AI community. This achievement not only highlights the model’s superior performance but also showcases Korea’s growing influence in the field of artificial intelligence.
EXAONE Deep offers valuable insights for industries seeking advanced AI solutions. Its proven capabilities in handling intricate mathematical problems, scientific reasoning, and coding tasks make it a versatile tool for businesses aiming to leverage AI for complex operational challenges. Continued advancements in models like EXAONE Deep are essential for driving innovation and maintaining competitiveness in an increasingly AI-driven world.
