Chinese AI startup DeepSeek has released an upgraded version of its large language model, DeepSeek-V3-0324, enhancing coding abilities, mathematical problem-solving, and reasoning skills. This latest iteration builds on DeepSeek V3, introducing additional parameters, optimized front-end web development capabilities, and improved proficiency in Chinese writing.
Performance and Benchmark Achievements
DeepSeek-V3-0324 has demonstrated superior performance across multiple benchmarks. It scored 59.4 on the American Invitational Mathematics Examination (AIME), a significant improvement over its predecessor’s 39.6. In coding, it achieved 49.2 on the LiveCodeBench benchmark, marking a 10-point increase. The model now has 685 billion parameters compared to DeepSeek V3’s 671 billion and operates under the MIT software license, making it widely accessible to developers on platforms like Hugging Face.
Real-World Applications and Expert Endorsements
Häme University of Applied Sciences lecturer Kuittinen Petri praised DeepSeek-V3-0324 for its ability to generate a fully functional, mobile-friendly website with 958 lines of code. Apple Machine Learning Research scientist Awni Hannun successfully ran the model on a 512-gigabyte M3 Ultra workstation, highlighting its efficiency. Mathematics Olympiad gold medallist Jasper Zhang tested the model on an AIME 2025 problem and confirmed its effectiveness.
Industry Impact and Cost-Effectiveness
DeepSeek’s success has fueled discussions on whether cutting-edge AI models can be developed at a fraction of the cost invested by US tech giants like OpenAI and Google. Its Mixture-of-Experts architecture enables scalable models with high performance at lower costs. This has also led to speculation about the upcoming release of DeepSeek R2, potentially arriving earlier than expected.
Competitive Edge and Market Influence
The launch of the R1 model in January demonstrated that competitive AI models could be developed at a lower cost, challenging OpenAI and Google. DeepSeek’s rise has influenced AI and tech-related stocks, leading to sell-offs as investors reassess the financial dynamics of AI development. Surpassing OpenAI’s ChatGPT, DeepSeek has become the most popular free app on Apple’s US App Store, reinforcing its competitive stance in the AI space.
Conclusion
DeepSeek’s latest advancements highlight the shifting AI landscape, where innovation is no longer confined to Silicon Valley. With efficient and cost-effective models, the company continues to push technological boundaries, positioning itself as a global AI leader.