Chinese artificial intelligence company that develops a large language model DeepSeek, has rolled out its latest AI model, DeepSeek-V3-0324, marking a significant leap in its capabilities and positioning itself as a formidable player in the industry.
Announcing the launch on X, the company wrote,
“DeepSeek-V3-0324 is out now!
Register for Tekedia Mini-MBA edition 17 (June 9 – Sept 6, 2025) today for early bird discounts. Do annual for access to Blucera.com.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register to become a better CEO or Director with Tekedia CEO & Director Program.
A major boost in reasoning performance,
Stronger front-end development skills,
Smarter tool-use capabilities. For non-complex reasoning tasks, we recommend using V3, just turn off “DeepThink”
API usage remains unchanged
Models are now released under the MIT License, just like DeepSeek-R1!”
This upgraded model has captured attention for its advancements in reasoning, coding, and multilingual proficiency. The company highlights that DeepSeek-V3-0324 outperforms its predecessor across multiple benchmarks, demonstrating notable gains in problem-solving and software development.
Key enhancements include
Superior Reasoning Ability
- MMLU-Pro: 81.2 (+5.3)
- GPQA: 68.4 (+9.3)
- AIME: 59.4 (+19.8)
Improved Code Generation
- LiveCodeBench: 49.2 (+10.0)
- More robust and executable front-end development
Enhanced Chinese Language Processing
- Refined writing style and content coherence
- Upgraded translation and text refinement
Advanced Function Calling
- Increased accuracy and reliability in API interactions
A standout improvement is the model’s performance in mathematical reasoning, particularly on the AIME exam, where it achieved a 59.4 score, a significant jump from 39.6 in the previous version. Additionally, its coding efficiency has been recognized with a 10-point increase on the LiveCodeBench evaluation.
With 685 billion parameters, DeepSeek-V3-0324 surpasses its earlier iteration while adopting an MIT license, making it more accessible for developers globally.
DeepSeek Positioning Itself as A Formidable Competitor to ChatGPT, Gemini, Anthropic, and Others
DeepSeek’s unveiling of V3-0324, marks a notable moment in the AI landscape, positioning the Chinese AI startup as a serious competitor against top players like ChatGPT, Gemini, Grok, and all other giant AI chatbot models.
This comes as some of these startups have continued to upgrade their AI models. Earlier this month OpenAI enabled Python-powered data analysis in ChatGPT. This enables users to ask the chatbot to perform tasks like running regressions on test data, visualizing complex business metrics, and conducting scenario-based simulations. Also just recently, it rolled out 4o image generation in ChatGPT and Sora to all Plus, Pro, Team, and Free users.
Elon Musk-owned xAI also added deeper search and image editing capabilities to the AI model Grok. The new “DeeperSearch” option works alongside the existing “DeepSearch” feature, though it takes more time to process queries. DeepSearch can answer questions about “latest AI news” in about a minute, pulling from 40 sources.
On the other hand, Google-owned Gemini, two days ago rolled out its latest AI model, Gemini 2.5 Pro, insisting on the model’s superiority as state-of-the-art technology across different debuts and benchmarks. The model notably ranked higher than xAI’s Grok 3 previews and OpenAI’s GOT 4.5 preview.
Despite all these updates, DeepSeek’s competitive edge is rooted in one very key factor, which is its cost-effective development and Operations. The Chinese startup has engineered its AI models to be both economical to train and efficient during inference. By implementing innovative engineering techniques, the company has significantly reduced inference costs while enhancing training efficiency. This approach allows DeepSeek to operate with lower expenses compared to many competitors, providing a substantial market advantage.
Looking Ahead
While OpenAI and Google currently lead the AI arms race, Deepseek’s emergence signals increased competition in the field. If the platform successfully delivers on its promises of superior performance and efficiency, it could disrupt the status quo and attract a loyal user base.
Notably, as AI technology continues to evolve, the battle for dominance in the AI market will intensify, offering consumers more choices and driving further advancements in the field.